Weekly Video Scripts

The 14 weekly build-video scripts · ~10 min each · PLACEHOLDER drafts. Every episode carries the temporary-by-design beat; day shot-lists link into each week's Shot list.

Atmosfarm — Weekly Build Video 01: Foundation / Ground

Build Week: 1 of 14 | Target runtime: ~10:00 | Status: PLACEHOLDER (draft — refine later)

Logline

Day one of one hundred. Before a single strut goes up, the whole project lives or dies on the ground beneath it — so this week is about getting the dirt level, the footprint marked, and every base anchor set true. Get this wrong and the dome fights you for the next three months.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): No slab, no concrete — screw piles and ground screws. The whole dome can be unbolted and lifted off this ground.

The rule behind every decision: temporary by design, and self-funded. This Kudla block is in a growth area — it'll likely be subdivided in 2–3 years — so I'm pouring nothing in; everything bolts, clamps and lifts and can move when the time comes. I'm funding it myself and building in public — no crowdfunding, no ask. One day Atmosfarm moves to a permanent property and goes in to stay, but that's a later chapter, not what this is about.

Cold open (0:00–0:30)

Visual: Drone "before" shot — slow rise over the [Adelaide Hills / South Australia] site at first light. Empty ground, the rough cleared footprint just visible. Hold on the emptiness for a beat. Then a fast timelapse of the footprint being cleared and the first peg going in.

VO (over the drone shot, quiet, honest):

"This is the ground. Right now there's nothing here. In a hundred days there's going to be a dome standing on it — one that pulls water out of the air. But none of that works if I get this part wrong."

VO (as timelapse kicks in):

"So before I build anything that goes up… I have to get everything that goes down — dead level."

Hard cut to title card.

Intro to-camera (0:30–~1:30)

Visual: Stephen to-camera, standing on the cleared site, tape measure or marking line in hand. Handheld, natural light. Site visible behind him.

To-camera (conversational, sets up the whole series):

"Okay — welcome to Atmosfarm. If you're new here: I'm Stephen, and I'm building a geodesic dome out here in [South Australia] over the next hundred days. It's a 6V dome — metal conduit frame — and the whole point of it is atmospheric water harvesting. It's going to pull drinking water straight out of the air."
"This is week one of fourteen build weeks. Every week I'll take you through what I actually got done — the wins, the screw-ups, all of it. There's also a livestream running every night on the [second channel] if you want to watch the unedited version while I'm covered in dirt."
"This week is the least glamorous and probably the most important: the foundation. The ground. Because a 6V dome has a LOT of base points, and every single one of them has to land exactly where the maths says it should. So — let's get the site set up."

On-screen text: Day 1 / 100 and Week 1 — Foundation

Build segments

Segment 1 — Clearing, leveling & setting the datum (1:30–4:00)

To-camera / VO:

"First job: clear it and level it. The datum is just my reference point — the one height everything else gets measured from. Once that's set, it doesn't move. Every footing, every anchor, gets referenced back to this."
"[I'm using a laser level / a string line and a water level] to find true level across the whole footprint. The site's got a bit of a fall on it [assumption: slight slope toward (direction)], so I need to know exactly how much I'm working against before I commit to anything."

Shot notes:

POV of clearing the footprint — boots, shovel, rake. Get the dirt and effort.
Close B-roll: laser level dot / string line catching the light, bubble vial settling.
Timelapse: the level sweep across the site.
Drone top-down to show the full cleared circle taking shape.

On-screen text: Set the datum — everything references this

Segment 2 — Marking the footprint & base ring (4:00–6:30)

To-camera / VO:

"Now the fun part — laying out the footprint. I find the centre, mark my radius, and snap the full circle. Then comes the part that takes patience: a 6V dome has [assumption: ~31] base vertices around that ring, and each one has a precise position. I'm not eyeballing any of these."
"Centre point, radius, angle, peg. Centre, radius, angle, peg. You do that wrong on point three and you don't find out until point twenty-eight won't close up."

Shot notes:

POV close-up: the pegging process — measure, mark, drive the peg, repeat.
Satisfying B-roll of pegs in a clean arc.
Drone reveal: the full pegged base ring from above — the circle finally readable as a dome footprint.
Timelapse of the whole marking-out sequence compressed to ~20 seconds.

On-screen text: 6V base ring — mark & anchor EVERY vertex

Segment 3 — Compaction, footings, anchors & sleeves (6:30–8:00)

To-camera / VO:

"Pegs mark it — but they don't hold it. So now I compact the ground at each point [assumption: plate compactor / hand tamp], set the footings, and put in the anchor points the frame will actually bolt down to."
"And here's the thing you have to do NOW or you never get to do it: anything that needs to run underground — the conduit sleeves, the drainage — goes in before anything covers it. Once the frame's up, this stuff is buried forever. Future me will be very angry at present me if I forget a sleeve."

Shot notes:

B-roll: compactor in action (good audio + dust).
Close-up: footing/anchor point being set and checked against the datum.
POV: laying conduit / drainage sleeves into trenches before backfill.
Mark each sleeve run on-screen with a simple graphic line.

On-screen text: Run sleeves & drainage BEFORE anything covers them

The science / why it matters (8:00–8:40)

Visual: To-camera, maybe crouched at one of the anchor points, or a simple overlay diagram of the dome on the footprint.

To-camera:

"Quick on why this matters so much. A geodesic dome spreads its load out across all those base points — so they all have to share it evenly. If one anchor sits high, or the ring isn't a true circle, the frame carries that error all the way to the top and the panels won't sit right."
"And on the water side — this dome's whole job is collecting moisture, so drainage isn't an afterthought. Where the water goes is designed in from day one, literally underground, right now."

On-screen text / graphic: simple diagram — load spreading to base ring + arrow showing drainage path.

Honest beat (8:40–9:10)

Visual: To-camera, end of day, tired light. Less polished. Maybe sitting on a material pile.

To-camera (real, unscripted feel):

"Honestly? Today there's nothing to show off. No dome. Just a circle of pegs and a level patch of dirt. [assumption: and one anchor point I already had to reset because I rushed it.] But this is the part nobody films and everyone regrets skipping. Slow today saves me weeks later. I think. We'll find out."

Outro + next-week tease (9:10–10:00)

Visual: Drone pull-up and away from the finished, pegged, level footprint — golden hour. The empty circle now clearly the start of something.

To-camera / VO:

"So that's week one. The ground's level, the footprint's marked, every anchor's in, and the sleeves are buried before anything covered them. The site is ready."
"Next week, it finally starts going vertical — I'm fabricating the dome struts. Cutting, drilling, flattening conduit, building the actual bones of this thing. That's where it starts to look like a dome."
"If you want the daily version, the livestream's on every night on the [second channel]. Subscribe so you don't miss week two — and I'll see you on day eight."

On-screen text: NEXT WEEK → Strut fabrication: the dome goes vertical + Subscribe + nightly livestream @atmosfarm

Shot list (quick reference)

Drone "before" — empty site at first light (cold open hero)
Timelapse — footprint clearing
To-camera intro — series setup, on the site
B-roll — laser level / string line, bubble vial settling
Timelapse — leveling sweep
Drone top-down — cleared circle
POV — pegging process (measure / mark / drive / repeat)
B-roll — pegs in clean arc
Drone reveal — full pegged base ring from above
B-roll — compactor running (dust + audio)
Close-up — footing/anchor set and checked to datum
POV — conduit/drainage sleeves into trench
To-camera — science beat (anchor point or with diagram overlay)
To-camera — honest end-of-day beat (tired light)
Drone pull-up/away — finished footprint at golden hour
[Night-work B-roll if any leveling/marking ran late — lighting is set up]

On-screen text / graphics

Day 1 / 100
Week 1 — Foundation
Set the datum — everything references this
6V base ring — mark & anchor EVERY vertex
Run sleeves & drainage BEFORE anything covers them
Diagram: load spreading across base ring
Diagram: underground drainage path
NEXT WEEK → Strut fabrication: the dome goes vertical
Subscribe + nightly livestream @atmosfarm
Lower-third (first appearance): Stephen — Atmosfarm

Runtime breakdown

Section	Start	Length
Cold open	0:00	0:30
Intro to-camera	0:30	1:00
Seg 1 — Level & datum	1:30	2:30
Seg 2 — Footprint & base ring	4:00	2:30
Seg 3 — Compaction / anchors / sleeves	6:30	1:30
Science / why it matters	8:00	0:40
Honest beat	8:40	0:30
Outro + next-week tease	9:10	0:50
Total		~10:00

↑ top

Atmosfarm — Weekly Build Video 02: Dome Frame — Fabrication

Build Week: 2 of [14] | Target runtime: ~10:00 | Status: PLACEHOLDER (draft VO + shot direction — refine later) First of 3 frame weeks. This week = pure fabrication. No raising yet.

Logline

The biggest job on the whole build starts here — turning a pile of raw metal conduit into every single strut of a 6V geodesic dome: cut, flattened, drilled, colour-coded and sorted. Get this right and the raise is easy. Get it wrong and nothing fits. So this week we build a production line.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): Every strut is cut and bolted, never welded shut — so the frame comes apart as easily as it goes together.

Atmosfarm is temporary by design and self-funded — this Kudla block sits in a growth area and will likely be subdivided in 2–3 years, so it's all built to come apart and move. (A permanent home is a someday thing, not today's story — no ask.)

Cold open (0:00–0:30)

Visual: Hard cut, no logo yet. Close macro on the chop saw biting into conduit — sparks, the scream of the cut, a strut dropping onto a growing pile. Quick-cut montage: flattening press slamming, drill bit punching through flattened end, a hand slapping a strip of coloured tape on a strut. Sound design forward, almost no talking.

VO (over the montage, dry and honest):

"A 6V dome has [hundreds] of struts. And they are not all the same length. Get one group wrong... and the whole thing won't close. So before I raise a single thing — I've gotta make every part. All of them. This week."

Visual: Snap to black on the last drill punch. Title card.

On-screen: ATMOSFARM / WEEK 2 — THE FRAME, PART 1: FABRICATION

Intro to-camera (0:30–~1:30)

Setting: Stephen standing in the build area, raw conduit bundles behind him, the empty dome footprint marked out on the ground. Daylight. Handheld, natural.

To-camera (conversational):

"Alright. Week two. Last week was [site + footing prep / the slab] — this week it gets real, because we start the frame. And honestly the frame is the biggest single job of the entire 100 days. It's so big I've split it into three weeks. This one — today — is fabrication. I'm not raising anything. I'm just making the parts."

"Here's the thing about a 6V dome. It's basically a sphere chopped into triangles, and to make a sphere out of straight bits of metal, the triangles can't all be identical. So the struts come in [several] different lengths — and they're really close to each other, like, [millimetres] apart. Close enough that if I eyeball it, I'll mix them up. Guaranteed."

"So the whole game this week is: cut every strut dead accurate, flatten and drill the ends so they bolt together, and then label everything so future-me at the raise knows exactly what's what. I'm setting up a little production line. Let's go."

Shot notes: One drone fly-over of the whole site drops in under the last line — establish the empty footprint vs the bundles of raw conduit, to sell the scale of "make all of this."

Build segments

Segment 1 — Cutting to length (~1:30–4:00)

Draft to-camera (at the saw):

"First station: cutting. I worked out the strut lengths from [the dome calculator / chord factors for a 6V] — there's [X] groups of lengths. I've written them on the board so I'm not second-guessing mid-cut."

"The trick is a stop-block on the saw. I set the stop once for a length, then I just feed conduit and cut, feed and cut — every strut in that group comes out identical without me measuring each one. Measure once, cut [forty] times. That's the whole idea of a jig."

VO over POV (GoPro / glasses cam):

"Same length, over and over, until the group's done. Then move the stop, new group, go again. It's repetitive — but repetitive is good. Repetitive means consistent."

Shot notes:

POV (GoPro/glasses cam) running the cut line — feed, clamp, cut, drop. Let it run real-time for a few cuts so the rhythm lands.
Macro B-roll: the stop-block, the pencil mark, sparks, offcut bin filling.
Timelapse: a pile of identical struts growing next to the saw.
On-screen callout of strut group + count as each group finishes [confirm exact numbers].

Segment 2 — Flattening + drilling the ends (~4:00–6:30)

Draft to-camera (at the press / vice):

"A round tube can't sit flat against another round tube. So both ends of every strut get flattened — squashed into a little paddle — and then drilled, so they can all bolt together at the hubs. This is the bit that's gonna eat my time, because it's [X hundred] struts times two ends times flatten-plus-drill."

"So again — jig. I made a [simple bending/flattening] setup and a drill jig so the hole lands in the same spot, the same distance from the end, every time. If those holes wander, the bolt holes won't line up at the raise and I'll be drilling on a ladder at the top of a half-built dome, which... no."

VO over POV + timelapse:

"Flatten. Flip. Flatten. Drill. Drill. Next. Once the line's dialled in you stop thinking and just feed it."

Shot notes:

POV through the flatten/drill line.
Macro: flattened paddle end, the drill jig registering the strut, swarf curling off the bit.
Timelapse of one full batch going through both stations.
Honest insert: [a flattened end that split / a hole drilled off-centre] — show the reject pile. Failures stay in.

Segment 3 — Colour-coding, labelling + sorting into rings (~6:30–8:00)

Draft to-camera (over the sorted struts):

"Last station, and honestly the most important one for keeping me sane: sorting. Every length group gets its own colour — [tape / paint mark] on the end — and a label. [Colour A] is one ring, [colour B] is the next, and so on, out to the [base ring]."

"When I raise this thing next week, I do not want to be measuring struts on the ground. I want to grab '[the red ones]', know exactly where they go, and bolt them up. The labelling I do now is what makes the raise fast and safe later."

Shot notes:

Top-down / overhead of struts laid out in neat colour-grouped bundles — visually satisfying, the payoff shot of the week.
B-roll: hand tabbing colour onto ends, writing the label, stacking bundles by ring.
Drone pull-up from the colour-sorted bundles to the empty footprint — "all of this is about to become that."
On-screen graphic: simple 6V diagram with rings colour-matched to the bundles [build later].

The science / why it matters (to-camera beat, ~8:00–8:45)

To-camera (calmer, explaining):

"Quick why-it-matters. A geodesic dome is strong because the load spreads through all these triangles instead of sitting on a few beams — that's why a frame this light can hold its own shape and [carry the harvesting gear]. But that only works if the geometry is right. The strut lengths aren't arbitrary, they're the math of fitting flat triangles onto a curved surface."

"So 'frequency' — the 6V — just means how finely the sphere's chopped up. Higher frequency, more triangles, rounder dome, but more different strut lengths to keep track of. That's the trade. More accuracy needed, smoother result. And that's exactly why this boring sorting week is the one that decides whether the raise goes smooth or turns into a nightmare."

Shot notes: Cut to the colour-coded diagram graphic; intercut the real sorted bundles so the abstract matches the physical.

Honest beat (~8:45–9:15)

To-camera (relaxed, end-of-day light or night):

"Real talk — this week is not glamorous. It's the same four moves a [thousand] times and my [arms / ears] are done. There were moments I wanted to skip the labelling and just crack on. But every dome build I've read about that went sideways went sideways here — sloppy parts. So I'm forcing myself to be slow and boring now, so I can be fast later. We'll see if that bet pays off next week."

Shot notes: Quiet handheld. Maybe a night shot under the work lights (night lighting's set up) — tired-but-satisfied, the day's finished bundles stacked behind him.

Outro + next-week tease (~9:15–10:00)

To-camera:

"So that's it — every strut for the whole dome, cut, flattened, drilled, colour-coded and stacked by ring. The hardest invisible week. Next week it finally goes vertical: we start the raise with the lower frame. That's where it stops being a pile of metal and starts looking like a dome."

"If you want to watch the unglamorous bits in full — the offcuts, the rejects, the swearing — that's all on the nightly livestream over on [the livestream channel], an hour every night while I build. Subscribe here so you don't miss the raise. See you next week."

Shot notes:

Final drone shot: orbit the empty footprint with the sorted bundles laid beside it.
Hard tease cut: 1–2s flash of [a strut going up / first hub bolted] for next week, then black.

On-screen: NEXT WEEK — THE RAISE BEGINS: LOWER FRAME / livestream channel handle / Subscribe.

Shot list (quick reference)

Drone: opening establish of site (empty footprint vs raw conduit bundles)
Drone: pull-up off colour-sorted bundles → footprint
Drone: closing orbit of footprint + finished bundles
POV (GoPro/glasses): cutting line — feed/clamp/cut/drop (real-time run)
POV: flatten + drill line
Macro: chop saw cut, sparks, stop-block, pencil mark
Macro: flattened paddle end, drill jig, swarf
Macro: colour tape onto ends, labelling, stacking
Timelapse: pile of identical struts growing
Timelapse: one batch through flatten + drill
Overhead / top-down: struts laid out in colour-grouped bundles (payoff shot)
Insert: reject pile (split end / off-centre hole)
To-camera: intro, segment pieces, science beat, honest beat, outro
Night: work-lights honest beat [optional]
Tease flash: strut going up / first hub [pull from Week 3 footage]

On-screen text / graphics

Title card: ATMOSFARM / WEEK 2 — THE FRAME, PART 1: FABRICATION
Per-group callout as each finishes: [GROUP / COLOUR] — [length] — [count] struts [confirm numbers]
Lower-third first time on camera: Stephen — building a water-harvesting dome in 100 days [confirm wording]
Diagram: 6V dome rings, colours matched to physical bundles [build in post]
Stat overlay: total struts / total cuts / total holes drilled this week [fill from real tally]
End cards: NEXT WEEK — THE RAISE BEGINS: LOWER FRAME, livestream handle, Subscribe

Runtime breakdown

Section	Start	Length
Cold open	0:00	0:30
Intro to-camera	0:30	1:00
Seg 1 — Cutting	1:30	2:30
Seg 2 — Flatten + drill	4:00	2:30
Seg 3 — Colour-code + sort	6:30	1:30
Science / why it matters	8:00	0:45
Honest beat	8:45	0:30
Outro + tease	9:15	0:45
Total		~10:00

PLACEHOLDER notes: confirm exact 6V strut group count, lengths, colour scheme and total strut tally; confirm Week 1 subject for intro callback; confirm livestream channel handle and lower-third wording. Bracketed items are assumptions to verify against the real build.

↑ top

Atmosfarm — Weekly Build Video 03: Dome Frame — Lower Frame

Build week: 3 of 14 (Frame, part 2 of 3) · Runtime: ~10 min · Status: PLACEHOLDER (draft to refine)

Logline

The frame leaves the ground. We raise the base row, bolt and lock the lower hubs, connect the lower rings — and for the first time, this thing actually looks like a dome. But before we go any higher, we stop and check level and square, because every error down here gets multiplied on the way up.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): Bolted hubs, not welds — this lower frame can be dismantled ring by ring and re-raised somewhere else.

Cold open (0:00–0:30)

Visual: Drone, low and slow, sweeping up from ground level past a half-raised triangle of conduit catching the late-afternoon light. Cut to a single hub being torqued — wrench, click, hands. Cut to the wide: the base ring standing on its own for the first time.

VO (over the top, quiet):

"Last week this was a pile of bent pipe on the ground. [pause] Today it stands up."

Visual: Quick beat — Stephen steps back, hands on hips, just looking at it. Hold half a second too long. Title card hits.

On-screen: ATMOSFARM / WEEK 3 — THE FRAME RISES

Intro to-camera (0:30–~1:30)

Visual: Stephen to-camera, base ring standing behind him, [morning light]. Handheld, slightly loose.

To-camera (conversational):

"Alright — week three. The frame.
Last week we cut and bent every strut and pre-assembled the hubs flat on the ground. This week the whole thing goes vertical. We raise the base row, we bolt the lower hubs down and lock them, and we tie in the lower rings.
And there's one job this week that matters more than anything else, and it's the least photogenic: checking that this lower frame is dead level and properly square before we build another inch on top of it. [beat] On a dome, a six-millimetre error down here turns into something way uglier by the time you reach the apex. Errors compound upward. So we measure twice, we measure again, and then we keep going.
It's also the first week this starts to look like the thing in my head. So — let's raise it."

On-screen: lower-third Stephen · Day [15] · Adelaide, SA

Build segments

Segment 1 — Raising the base row (1:30–4:00)

To-camera / VO (mixed):

"The base row is the foundation of everything above it. Every strut up there eventually transfers load back down into this ring, so it has to sit true.
[I'm doing this with two people and a lot of patience / solo with props — confirm crew.] We walk each section up, prop it, and only loosely pin the hubs at first — nothing torqued yet. You want the whole base ring standing and roughly round before you commit any single joint, otherwise you fight yourself."

Shot notes:

POV (chest mount): lifting and walking a base section up — hands, conduit, the strain. Raw audio.
B-roll: props and temporary bracing going in; the ring slowly closing into a circle.
Timelapse: the full base row going from flat to standing — this is a key timelapse, lock the camera off on a tripod and don't move it.
Drone: one orbit at base-ring height once it's standing, to read the circle from above.

Segment 2 — Bolting and locking the lower hubs (4:00–6:30)

To-camera (the hub explainer):

"Okay, let me show you how a hub actually works, because this is the heart of a strut dome.
[Hold a hub up to camera.] Every joint where struts meet is a hub. The struts flatten at the ends, they overlap here, and they get bolted through a common point. The geometry of the dome — the angles, the ring layout — all of that is baked into how these flattened ends are drilled. [reference: this is the 6V conduit dome layout — see build notes.]
On the lower ring, once everything's lined up and the frame's reading round and level, I go around and torque each one properly and lock it — [thread locker / lock nuts / both — confirm]. Loose hub now equals a wobble later. So this is where the frame stops being a loose sketch and becomes a structure."

Shot notes:

POV (close, through the hub): the overlap, the bolt going through, the wrench. Macro if the lens allows.
Insert: clean static shot of one finished, locked hub. This doubles as a graphics plate.
B-roll: moving methodically around the ring, hub to hub. Walk-and-work rhythm.
Cutaway: [night-work option] — if this runs into the evening, grab the work lights coming on and a hub being torqued under them. Good texture, ties to the livestream.

Segment 3 — Connecting the lower rings + the level/square check (6:30–8:30)

To-camera / VO:

"With the base locked, the lower rings tie the base verticals together and start pulling the shape inward — you can suddenly see the dome curving.
And here's the part I promised wasn't pretty. [walk to it] Before I go one ring higher, I check it. Level around the base — [laser level / water level / spirit level — confirm tool]. Then square, or as close to square as a circle gets: I'm checking that opposing points are equidistant and the ring hasn't gone egg-shaped. [hold up the tape] If something's out, NOW is when it's cheap to fix. In two weeks it's a nightmare.
[Honest aside if it applies: 'this corner was out by [X] and I had to back off three hubs to bring it in.']"

Shot notes:

B-roll: tape measure across the diameter; level on the ring; Stephen crouched, checking, adjusting.
Insert: the measurement readout / level bubble — get the actual number on screen.
Drone — HERO TIMELAPSE: the full week compressed — flat pipe to a standing, ringed lower frame. This is the money shot of the episode. Fixed drone position or locked high tripod so the frame "grows" in place.

The science / why it matters (8:30–9:10)

VO over graphics:

"Why obsess over level and square? A geodesic dome is strong because load spreads evenly through the whole network of struts — no single member carries much on its own. But that only works if the geometry is true. Build the base out of square and you force struts higher up into angles they were never cut for. The dome fights itself: stressed joints, gaps at the apex, a frame that won't close. [confirm phrasing with build notes.]
Get the bottom right, and the top almost builds itself."

On-screen / graphics: simple animated diagram — even load distribution across a true dome vs. a skewed base pushing stress into upper hubs. [GRAPHIC TBD]

Honest beat (9:10–9:30)

To-camera (quiet, real):

"Honestly? [The base ring fought me / I under-ordered bolts / the level check cost me half a day — confirm what actually happened.] [beat]
But it's standing. And for the first time I walked up to it this afternoon and it looked like a dome instead of an idea. [small laugh] That's a good day."

Shot notes: golden-hour, Stephen and the standing frame, drone pulling back slowly. Let it breathe.

Outro + next-week tease (9:30–10:00)

To-camera:

"Next week we go all the way up — closing the upper rings and pulling everything in to the apex. The top of the dome. That's the week it either closes clean [or it doesn't].
If you want to watch it happen in real time, the build's live every night over on [livestream channel name] — an hour, every evening, work lights and all.
Subscribe, and I'll see you at the top. Cheers."

Visual: final drone rise off the standing lower frame, tilting up toward the empty sky where the apex will be. Hard cut to black.

On-screen: NEXT WEEK — CLOSING TO THE APEX / Live every night · [channel] / subscribe prompt

Shot list

Cold open drone rise past half-raised frame (golden hour)
Hub torque close-up (cold open + segment 2)
Wide: base ring standing on its own
Intro to-camera, frame behind (morning)
POV chest-mount: walking base sections up
Temp bracing / props going in (B-roll)
Timelapse: base row flat → standing (locked tripod)
Drone orbit at base-ring height
To-camera hub explainer with hub in hand
POV through-the-hub bolt + wrench (macro)
Static insert: one finished locked hub (graphics plate)
Hub-to-hub walk-and-work B-roll
[Night-work option] hubs under work lights
Lower rings tying in, shape curving inward
Tape-across-diameter + level on ring (B-roll)
Insert: measurement / bubble readout (real number)
HERO TIMELAPSE: full week, flat pipe → standing ringed frame (fixed position)
Science graphic plate (even vs skewed load)
Honest beat, golden hour, drone pull-back
Outro drone rise tilting to empty apex sky

On-screen text / graphics

ATMOSFARM / WEEK 3 — THE FRAME RISES (title)
Lower-third: Stephen · Day [15] · Adelaide, SA
Label callout: LOWER HUB (over hub explainer)
Measurement overlays: level reading + diameter check (real numbers)
Science diagram: true dome vs skewed base [GRAPHIC TBD]
NEXT WEEK — CLOSING TO THE APEX
Live every night · [livestream channel]
Subscribe prompt (end)
Safety note overlay (segment 1/2): Build is now vertical — props, footing, no one under a raised section [confirm wording]

Runtime breakdown

Section	Start	Duration
Cold open	0:00	0:30
Intro to-camera	0:30	1:00
Seg 1 — Raising the base row	1:30	2:30
Seg 2 — Bolting / locking lower hubs	4:00	2:30
Seg 3 — Lower rings + level/square check	6:30	2:00
The science / why it matters	8:30	0:40
Honest beat	9:10	0:20
Outro + next-week tease	9:30	0:30
Total		10:00

↑ top

Atmosfarm — Weekly Build Video 04: Dome Frame — Upper Frame & Apex

Build week: Week 4 of 14 · Runtime: ~10 min · Status: PLACEHOLDER (draft VO/to-camera + shot direction — refine later)

Logline

The third and final frame week — the payoff. The upper rings close in, the apex hub locks home, and a pile of conduit finally becomes a dome. Then the unglamorous part: bracing it out and torque-checking every fixing for the load the skin will throw at it. The whole week happens at height, and that's the real story.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): The whole frame is a bolt-together kit — apex down to base, it all comes apart for the move.

Cold open (0:00–0:30)

Visual: Hero drone shot. Start low, behind the dome, then climb and arc over the top — the frame complete for the first time, full triangulated lattice against the [late-afternoon] sky. Hold the apex in centre frame as the drone passes directly over.

VO (over the shot, quiet, half-laughing):

"Three weeks ago this was a stack of bent pipe on the grass. Today... yeah. It's a dome."

Cut to: Quick POV looking straight up from the ground through the inside of the frame, sun flaring through the triangles.

On-screen text: WEEK 4 — THE DOME CLOSES

Intro to-camera (0:30–~1:30)

Visual: Stephen to-camera, standing inside the frame, can put a hand on a strut. Frame fills the background. [Golden hour if we can time it.]

To-camera (conversational):

"Alright — week four. This is the last of the three frame weeks, and it's the one I've been a bit nervous and a bit excited about in equal measure.
First three weeks we did the base, the lower rings, the middle. This week we close it off — the top rings go in, and then the apex hub. That's the single piece that ties the whole top of the dome together. Once that's in, it's a complete structure. Done. A self-supporting dome.
But here's the honest bit: everything this week happens up high. I'm not working off the ground anymore. So as much as this is the fun week, it's also the week I have to be the most switched on. We'll talk about that.
Let's get the top on."

On-screen text: Frame week 3 of 3 · Upper rings → apex → brace → torque check

Build segments

Segment 1 — Closing the upper rings (1:30–4:00)

To-camera / VO (mix):

"The geometry tightens right up near the top. Down low the triangles are big and forgiving. Up here the struts are shorter, the angles are sharper, and there's a lot less room to get your hands in. Every strut I add now is pulling the whole top of the dome in toward a point."

VO over timelapse:

"I'm working ring by ring — get a full ring of struts loosely bolted, check it's sitting true all the way around, then come back and bring everything up to tension. If I tighten as I go, one side pulls ahead of the other and the whole thing goes lopsided on me. Ask me how I know. [ref earlier week if it happened]"

Honest micro-beat:

"[PLACEHOLDER — slot in a real snag here: a hole that didn't line up, a strut cut [Xmm] short, having to go back down for a different bolt. Keep it real.]"

Shot notes:

Timelapse: one full upper ring going in, locked-off camera on tripod [or clamped to scaffold].
POV/chesty: hands bolting a strut at height, looking down through the frame.
Drone: slow orbit at apex height watching the top close in.
B-roll: macro of a hub plate, bolts, the spacing between triangles tightening.
Detail: spirit level / [digital angle finder] on a strut checking true.

Segment 2 — The apex hub goes in (4:00–6:00)

This is the hero moment of the episode. Give it room.

To-camera (before):

"This is it. This is the piece. Every strut on the top half of this dome lands on this one hub. If this sits right, the dome's geometry is locked. If it's off, I feel it everywhere. No pressure."

VO over the POV:

"Lining it up is the fiddly part — [six] struts all wanting to meet in the same place at the same time, and gravity not helping. [Note method: temporary support pole / strap / second person handing struts up.]"

The moment — minimal VO, let it breathe:

"...and that's the last one. That's the apex in."

Visual: POV looking straight down into the apex hub as the final fixing seats and gets torqued. Then cut wide.

To-camera (immediately after, genuine):

"[laughs / exhales] That's a dome. That is officially a dome."

Shot notes:

POV/chesty cam: THE shot — final hub fixing going in, looking down into the hub.
Second angle: locked-off wide catching the moment the apex seats (multi-cam this — don't rely on one take).
Drone: pull straight up off the apex the instant it's in — reveal the completed top.
B-roll: the apex hub from below, struts radiating out like spokes.
Reaction: keep the camera rolling on Stephen — the real reaction is the gold.

Segment 3 — Brace it & torque-check everything (6:00–8:00)

To-camera:

"Now the part nobody puts in the highlight reel. A frame that's just bolted together is not finished. I've got to brace it out so it can't rack — can't twist — and then I go around and torque-check every single fixing on the dome. Every one.
Because in a few weeks this frame carries the skin, and the skin catches wind. That's real load. A loose bolt I shrug at today is a failure point later. So this is the boring, methodical, do-it-properly part."

VO over timelapse/montage:

"[Bracing method PLACEHOLDER — diagonal bracing / cross-ties / ground anchors, as designed.] Then the torque wrench comes out, set to [spec — confirm value], and I work around the dome in a [systematic pattern] so I don't lose track of which ones I've done. [Method to mark checked fixings: paint pen / tape mark.]"

Shot notes:

Timelapse: the full torque-check lap around and up the dome.
Detail/macro: torque wrench clicking off on a fixing, paint-pen mark going on.
POV: bracing being added.
Drone: final orbit now that it's braced and checked — the frame at rest.
On-screen graphic: simple count-up of fixings checked [confirm total].

The science / why it matters

VO over [simple animation / overlay graphic + B-roll]:

"Quick why-it-matters. A geodesic dome is strong because it's a triangulated shell — load on any one point gets shared out across the whole structure instead of dumping onto a single joint. That's the whole trick. But it only works if it's actually complete and actually tight. An open frame, or a frame full of loose bolts, doesn't behave like a shell — it behaves like a bunch of separate sticks.
Closing the apex is what turns it from sticks into a shell. The brace and the torque-check are what keep it one. That's why this week mattered more than it looks."

On-screen text/graphic: Triangulated shell = load shared, not concentrated · [6V dome · X struts · X hubs — confirm]

Honest beat (8:00–8:45)

Visual: Stephen sitting on the ground [or a crate], dome behind him, light dropping. Tired-honest, not performed.

To-camera:

"I'll be straight with you. The thing I keep coming back to this week isn't how it looks — it's the height. I spent the whole week up off the ground, hands full, leaning into a structure I was still putting together. That's the bit you have to respect.
So: [PLACEHOLDER — name the actual safety setup. Harness / anchor points / scaffold / not working alone / not working tired / the night-work lighting so I'm never up there squinting in the dark]. None of that's optional. The dome's not worth getting hurt over, and a video definitely isn't.
[Optional real wobble: a moment I didn't love up there, or a thing I changed mid-week to make it safer.]
But — a third of the hardest job on this whole build is done. The frame is up. That's a good day."

Outro + next-week tease (8:45–~10:00)

Visual: Back to a hero drone shot — completed, braced frame, [dusk / lights coming on]. If night: the dome lit from within by the night-work rig.

To-camera / VO:

"That's week four. The dome is a dome. Frame's complete, braced, every fixing checked and marked.
Next week we stop building up and start building in — the air intake structure. This is the front end of the whole water-harvesting idea: how the dome actually pulls air in so we can pull water out of it. It's where this stops being a cool-looking frame and starts being a machine.
If you want the unedited version of all of this — the mistakes, the do-overs, the real-time problem solving — I'm live an hour every night on [livestream channel name / link]. Come hang out.
See you next week."

On-screen text: NEXT WEEK — THE AIR INTAKE · Live every night → [channel] · Atmosfarm · @atmosfarm

Shot list

Hero / drone

Cold-open fly-over of completed frame (climb + arc over apex) — HERO
Slow orbit at apex height during upper-ring close-in
Vertical pull-up off the apex the instant the hub seats
Final braced-frame orbit (frame at rest)
Dusk / night-lit hero for outro

POV / chesty

Bolting struts at height, looking down through frame
THE apex shot — final hub fixing going in, looking into hub
Bracing being added
Looking straight up through the inside of the frame (ground)

To-camera

Intro inside the frame
Pre-apex ("this is the piece")
Post-apex genuine reaction (keep rolling)
Honest beat — safety / height, seated, low light

B-roll / detail / macro

Hub plates, bolts, tightening triangle spacing
Apex hub from below — struts radiating
Torque wrench clicking off + paint-pen mark
Level / angle finder checking strut true

Timelapse

One full upper ring going in
Torque-check lap around/up the dome

Safety

Real footage of harness / anchors / scaffold / lighting rig in use [confirm setup]

On-screen text / graphics

Cue	Text	When
Title card	`WEEK 4 — THE DOME CLOSES`	Cold open
Sub	`Frame week 3 of 3 · Upper rings → apex → brace → torque check`	Intro
Lower third	`Stephen James Lovegrove · Atmosfarm · Adelaide, SA`	First to-camera [confirm]
Graphic	`Triangulated shell = load shared, not concentrated`	Science segment
Spec	`6V dome · [X] struts · [X] hubs`	Science segment [confirm numbers]
Counter	`[X] fixings checked`	Torque-check segment [confirm total]
Tease	`NEXT WEEK — THE AIR INTAKE`	Outro
Persistent	`Live every night → [livestream channel/link]`	Outro
End card	`Atmosfarm · @atmosfarm`	End

Runtime breakdown

#	Section	Start	Length
1	Cold open (drone fly-over)	0:00	0:30
2	Intro to-camera	0:30	1:00
3	Segment 1 — Upper rings close in	1:30	2:30
4	Segment 2 — Apex hub goes in (hero)	4:00	2:00
5	Segment 3 — Brace & torque-check	6:00	2:00
6	The science / why it matters	8:00	0:00*
7	Honest beat (safety / height)	8:00	0:45
8	Outro + next-week tease	8:45	1:15
	Total		~10:00

\*Science beat is woven into Segment 3's tail / runs under B-roll — budget ~0:30 inside the 6:00–8:00 block rather than as a separate stop. Adjust in edit. [Refine timings against real footage.]

↑ top

Atmosfarm — Weekly Build Video 05: Intake

Build week: Week 5 of [14] · Runtime: ~10 min · Status: PLACEHOLDER (draft VO/to-camera + shot direction — refine later)

Logline

This week the dome starts breathing. We build and aim the air intake at the prevailing wind, lock it down against gusts, screen it against bugs and leaves, and run the first airflow test — the moment we find out whether this thing actually pulls air the way the whole water-from-air system needs it to.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): The intake ducting clamps on — unclip it and it travels with the rig.

Cold open (0:00–0:30)

Shot notes: Tight, textured opening. Smoke (or a couple of fluttering streamers [assumption]) tied off near the unfinished intake mouth, drifting. Slow push-in. Ambient wind, no music yet. Cut to Stephen's hand holding an anemometer [assumption], the little cups spinning up.

VO (Stephen, low/conversational):

"Everything we're building — every panel, every litre of water this dome is ever going to make — it all starts with moving air through one hole. (beat) So today the whole project kind of comes down to one question: does it actually pull?"

On-screen: Streamers snap and hold in one direction. Hard cut to black.

Title card: ATMOSFARM — WEEK 5: INTAKE

Intro to-camera (0:30–~1:30)

Shot notes: Stephen to-camera, standing at the marked intake position on the dome, wind visibly moving his shirt/hair (sell the wind). Handheld, slightly loose. Dome in frame behind him.

To-camera (natural, can ramble a touch):

"Alright — week five. If you're new here: I'm Stephen, this is Atmosfarm, and we're [100] days into building a geodesic dome out in [South Australia] that's designed to harvest water straight out of the air. Drinking water, out of the sky, no rain required.
Up to now we've basically been building the skeleton — the frame, the hubs, the skin. This week we start building the lungs. The intake. This is the mouth of the whole thing, where outside air gets sucked in before it ever hits the water-2-air gear we'll install later.
And here's the bit people don't think about: if I put this in the wrong spot, or point it the wrong way, the dome fights itself for the next [decade]. Get the intake right and the wind does half my work for me, for free, forever. So no pressure. (laughs) Let's go figure out which way the wind blows — literally."

On-screen text: "Week 5 · The Intake · Day [29]–[35]"

Build segments

Segment 1 — Confirm prevailing wind + mark the position (1:30–3:30)

To-camera / VO mix:

"First job — don't trust a vibe, trust the data. I pulled the long-term wind records for [the site / nearest BOM station] and the prevailing wind here is out of the [SW] for most of the year [assumption — confirm with local BOM data]. But records lie a little once you've got buildings, trees, and a big round dome messing with the airflow. So I'm also doing the cheap version: streamers on stakes, a few mornings running, and just… watching."

Shot notes:

Screen-grab / clean graphic of a wind rose [assumption] over the site.
Time-lapse of streamers on stakes across different times of day.
Stephen with a compass / phone app confirming bearing, then marking the intake position on the frame with paint pen / tape.
Drone up-and-over to show the marked spot relative to the dome and the open ground the wind crosses.

VO:

"Records and streamers agree — [southwest]. So that's where the mouth goes. I'm marking it [low on the windward face / at the X panel] so it grabs clean, unobstructed air before anything else can steal it."

Segment 2 — Build + mount the intake structure (3:30–6:00)

To-camera (workshop / on-frame):

"The intake itself is [a framed mouth feeding into ducting] — built to match the dome's metal-conduit language so it doesn't look bolted-on. [Material assumption: conduit + sheet / ply throat + flexible or rigid ducting — confirm spec.] Two things matter here and only two: it has to point exactly into that wind, and it has to NOT rip off in a gust."

Shot notes:

Build montage: cutting, drilling, test-fitting the intake frame. Real workshop sound.
Close-ups of the throat/funnel shape.
Mounting to the dome frame — show the actual fixings.
A genuine fiddly moment (test-fit doesn't sit flush first go) — keep it in.

VO over montage:

"I'm flaring the mouth a little — wider opening necking down into the duct — so it scoops more air than the duct diameter alone would. [Assumption — validate in the airflow test.] Whether that actually helps or just looks clever, we'll know in about an hour."

Wind-load beat — to-camera:

"This is the part I keep going back to check. A flat-ish surface aimed straight into the prevailing wind is, structurally, a sail. So I'm bracing it [bracket: back-stays / gusset plates / through-bolts into the hub] and I'd genuinely rather over-build this than watch it cartwheel across the paddock on the first big blow."

Segment 3 — Mesh screen + stub the ducting (6:00–7:15)

To-camera / VO:

"Two finishing jobs. One — a mesh screen over the mouth. Bugs, leaves, the occasional confused bird — none of that gets into the system. [Mesh spec assumption: stainless insect-grade — fine enough to stop debris, open enough not to choke airflow.] It's a balance; too fine and I strangle the very draw I just built.
Two — I'm routing the ducting and leaving it stubbed [here], capped off, ready for the water-2-air tie-in in a few weeks. Future me will thank present me. Probably."

Shot notes:

Cutting and fitting the mesh, close-up of the weave.
Following the duct run along the frame to the capped stub.
Tap the cap, point at it: "that's a problem for week [9]."

The science / why it matters

Shot notes: Stephen to-camera, or VO over a simple animated cutaway [assumption — needs motion graphic]: air enters intake → passes the water-2-air unit → moisture pulled out → routed to downpipe/storage.

VO:

"Quick why-this-matters. Atmospheric water harvesting is basically: pull in humid air, cool it past its dew point, collect the condensation. Every step downstream depends on volume of air through the system. More clean airflow in, more water out. Full stop.
That's why the intake isn't a detail — it's the throttle for the entire dome. Point it into the prevailing wind and the wind itself pre-feeds my system before any fan or [pump] even spins up. It's the closest thing this build has to free energy. So yeah. One hole. Whole project. (beat) Let's test it."

On-screen text: "More airflow in → more water out."

Prototype test (7:15–8:45) — FIRST AIRFLOW TEST

Shot notes: This is the episode's payoff. Film it like a reveal. Smoke source / streamers at the mouth, anemometer [assumption] reading airflow. Get a clean readout shot. Both a calm-ish moment and a gusty moment if the day allows.

To-camera (genuine, unscripted energy):

"Okay. First airflow test. No fan yet — this is purely: does the placement and the wind do the job on their own. I've got smoke at the mouth so we can actually see the draw, and the anemometer for a number. [Assumption: anemometer available; otherwise streamer + visual only.]"

During test — react live:

"There — you see that? It's pulling. The smoke's getting dragged straight in… [reading X m/s — placeholder, fill real number]. That's [better / about / less] than I hoped."

Honest possibility — leave room for either outcome:

"[If it underperforms:] Honestly that's softer than I wanted. Could be the mesh, could be the angle, could be I'm sheltered by [the shed]. That's the test doing its job — better to find out now than after I've plumbed the whole system into it."

On-screen text: "FIRST AIRFLOW TEST · [X.X] m/s [placeholder]"

Honest beat (8:45–9:15)

Shot notes: Quieter. Maybe sitting on the frame near the intake, end of day, light going. Drop the energy, be real.

To-camera:

"I'll be straight with you — I second-guessed the placement [three times] today. There's a version of this where the prevailing wind shifts seasonally and I've optimised for the wrong [six months]. I don't fully know yet. (beat) But that's the deal with building something nobody's handed you a manual for. You make the best call with the data you've got, you test it, and you stay honest when it's not perfect. The intake's in. It pulls. We move."

Outro + next-week tease (9:15–10:00)

Shot notes: Stand up, back to wide. Drone weekly progress shot — slow rise revealing the dome with the new intake, the full site, the marked wind direction. This is the recurring weekly hero shot.

To-camera / VO:

"That's the lungs in. The dome's breathing for the first time — and it's breathing the right way.
Next week we build the downpipe support frame — the bones that'll carry the water once we actually start making it. Getting closer.
If you want the unfiltered version, I'm live every night on [the livestream channel] — building, failing, fixing, the whole mess in real time. Come hang out. And subscribe so you don't miss the next one. Cheers."

On-screen text:

"NEXT WEEK · Week 6 · Downpipe Support Frame"
"LIVE every night → [livestream channel handle]"
"@atmosfarm"

Shot list

Cold open: smoke/streamers at unfinished intake mouth, slow push-in
Anemometer cups spinning, close-up [assumption]
Title card insert
Intro to-camera at marked intake position, wind visible
Wind rose / BOM data graphic [assumption]
Streamers-on-stakes time-lapse, multiple times of day
Compass/phone bearing confirmation + marking the position
Drone: over-the-shoulder of marked spot vs dome + open ground
Intake build montage (cut, drill, test-fit) with real audio
Throat/funnel flare close-ups
Mounting + fixings detail
Failed/fiddly test-fit moment (keep authentic)
Wind-load bracing to-camera + bracing detail shots
Mesh screen fitting + weave close-up
Duct run follow to capped stub
Science cutaway animation [assumption — motion graphic needed]
PROTOTYPE: first airflow test — smoke draw + anemometer readout (calm + gusty)
Honest beat: end-of-day on the frame, fading light
Drone weekly progress hero shot (recurring) — rise reveal of dome + intake
Outro to-camera + walk to wide
NIGHT [optional]: any catch-up build under the night-work lighting rig

On-screen text / graphics

ATMOSFARM — WEEK 5: INTAKE (title card)
"Week 5 · The Intake · Day [29]–[35]"
Wind rose graphic — prevailing direction [SW, confirm]
"More airflow in → more water out."
"FIRST AIRFLOW TEST · [X.X] m/s" (placeholder reading)
Lower-third: mesh spec / material callouts [assumption]
"NEXT WEEK · Week 6 · Downpipe Support Frame"
"LIVE every night → [livestream channel handle]"
"@atmosfarm"

Runtime breakdown

Section	Start	Duration
Cold open	0:00	0:30
Intro to-camera	0:30	1:00
Seg 1 — Wind + mark position	1:30	2:00
Seg 2 — Build + mount intake	3:30	2:30
Seg 3 — Mesh + stub ducting	6:00	1:15
Science / why it matters	7:15	— (folded into pacing; see note)
Prototype — first airflow test	7:15	1:30
Honest beat	8:45	0:30
Outro + next-week tease	9:15	0:45
Total		~10:00

Note: "The science / why it matters" runs ~0:45 and is intended to sit just before the test (around 6:30–7:15) as the lead-in — trim Seg 3 or Seg 2 montage slightly to absorb it and keep the total at ~10:00. Adjust in edit.

↑ top

Atmosfarm — Weekly Build Video 06: Downpipe Frame

Build Week: 6 of 14 | Runtime: ~10 min | Status: PLACEHOLDER (draft VO / to-camera + shot direction — refine later)

Logline

The frame that everything downstream hangs off goes up this week — footings, plumb uprights, and cross-bracing until it doesn't budge under hand-load. And in a quiet corner of the build, we kick off the bio-filter four weeks early, because biology runs on its own clock.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): Bolted, braced and free-standing — no footings poured, so this frame lifts straight out.

Cold open (0:00–0:30)

Shot: Tight, low POV on a gloved hand giving a finished upright a hard two-handed shove. The frame doesn't move. Hold the stillness for a beat.

VO (over the shove, dry):

Everything I build after this week hangs off this frame. The downpipe, the flow into the filter, the water that ends up in the pond. So if this thing wobbles — all of that wobbles.

Shot: Snap-cut montage — a footing going in, a level bubble settling dead-centre, a brace being clamped, the drone lifting off. Fast, 3–4 frames each, cut to the music hit.

VO:

So this week I'm not building anything you can see water move through yet. I'm building the thing that holds it all up. [Plus a job I have to start now or I'm stuffed in four weeks.]

On-screen text: WEEK 6 — DOWNPIPE FRAME

Intro to-camera (0:30–~1:30)

Shot: Stephen to-camera on site, frame footprint marked out behind him (string lines / marking paint visible). Handheld, natural light, conversational.

To-camera (draft):

Morning. So — week six. Last few weeks we've been [up high on the dome / sorting the catchment surface — adjust to actual prior weeks], and this week the job drops down to ground level. We're building the support frame for the downpipe run.
Here's why this one matters more than it looks. The downpipe is what carries the water off the dome and down into the filtration. The frame I'm building today is what carries the downpipe. So the geometry of this — the heights, the angles, how rigid it is — that decides whether water actually flows the way I want it to, or just sits and pools and goes nowhere.
Two jobs today. One: get this frame up, plumb, and braced so solid I can hang my whole body off it. Two — and this is the one nobody expects in a "build the frame" episode — I'm starting the bio-filter today. Four weeks before the pond even gets water in it. I'll explain why when we get there, because the reason's actually pretty cool.

On-screen text: TODAY: 1. Frame up + braced 2. Seed the bio-filter (early)

Build segment 1 — Layout, footings & fixings (1:30–~3:30)

To-camera / VO (draft):

First thing — and this is the boring bit that everyone skips and everyone regrets — I'm setting out the layout. Where every upright lands. Get this wrong by a couple of centimetres now and the downpipe's fighting me the whole way down.

Shot notes:

Overhead / high-angle of string lines and marking paint defining the frame footprint.
POV down the run, sighting along the line to show it's straight.
Footings going in — [concrete footing pads / driven ground anchors / bolted base plates — confirm actual method]. Hands-in-frame, real work, no narration over the satisfying bits.

VO over footing work:

Everything load-bearing has to land on something that won't shift. So the footings go in first, levelled off each other, [set and left to cure if concrete — note timing]. The fixings get set while I've got clear access — way easier now than reaching past a built frame later.

On-screen text: Footings + fixings — set before anything goes vertical

Build segment 2 — Uprights up & plumb (~3:30–5:30)

To-camera / VO (draft):

Now the uprights. The whole game here is one word: plumb. Dead vertical. I'm checking every single one with the level on two faces, because a post that's leaning even slightly throws the whole downpipe line off — and you can't see it by eye until it's too late.

Shot notes:

POV raising the first upright, getting it roughly held [temporary prop / clamp].
Macro on the spirit level bubble settling between the lines. Hold it. This is the money shot for "did he do it right."
A wide showing the uprights standing but not yet braced — deliberately a bit gawky, so the bracing payoff lands next.

VO:

Right now these are standing but they're not trustworthy. A frame of bare uprights is a wobbly thing. It's the bracing that turns a row of posts into a structure — and that's next.

On-screen text: Plumb on two faces — every upright, no exceptions

Build segment 3 — Cross-bracing until it's rigid (~5:30–7:30)

To-camera / VO (draft):

This is the part that actually does the work. Cross-bracing. The uprights handle the weight going down; the bracing handles everything trying to push it sideways — wind, the weight of a full downpipe, me leaning on it. Triangles. Triangles are what stop a square frame folding over like a cardboard box.

Shot notes:

POV / over-shoulder fitting the diagonal braces, clamping, fixing.
B-roll of the geometry forming — the open rectangles becoming braced triangles. Let the shape tell the story.
The hand-load test (callback to cold open): both hands, full shove, frame holds. Maybe a second angle of it.

VO over the test:

The test isn't a number, it's a feel. I lean on it, I shake it, I basically try to wreck it. If it gives, I add more bracing until it doesn't. [Add brace count / final config once known.]

To-camera (after the test, honest):

Solid. That's the kind of solid where I'll happily hang a downpipe full of water off it and not think about it again.

Sub-segment — position the mounting points:

VO:

Last thing on the frame — I'm marking and fixing the downpipe mounting points at the right heights. This is where the geometry pays off: the mounts have to step down at [the correct fall / gradient] so the water actually runs downhill into the filter instead of sitting flat. Measured off the footings, not off the ground — the ground's never as level as it looks.

On-screen text: Bracing = triangles = rigid / Mounts set to fall — water has to run downhill

Build segment 4 — Seeding the bio-filter (early) (~7:30–8:30)

To-camera (draft, the "why now" beat):

Okay. This is the one that doesn't belong in a frame episode — except it completely does. I'm seeding the bio-filter media today. Four weeks before the pond fills in week ten.
Here's the thing people don't get about biofiltration: the filter media doesn't clean water on day one. It has to grow a living film — a colony of bacteria — across all that surface area. That's what actually breaks down the gunk. And you can't rush bacteria. They colonise and cycle on their own timeline — call it [~4 weeks] to mature.
So if I waited until the pond was ready in week ten and then dropped the media in, I'd have a filter full of dead rock doing nothing for a month. By starting it now, by the time water's flowing, the biology's already alive and ready to work.

Shot notes:

Close-up of the filter media [type — bio-balls / lava rock / K1 media — confirm] going into [the filter vessel / tub].
POV adding [the seed culture / starter — confirm method: dechlorinated water + ammonia source / starter bottle / pond seed].
A small label/marker with the start date being written — visual proof of the four-week head start.

VO:

I'll keep it [aerated / circulating / fed — confirm] and check on it through the livestream over the next few weeks. By week ten it should be a living, breathing filter.

On-screen text: Bio-filter seeded: [DATE] — cycling now, ready by Week 10 / Bacteria don't hurry

The science / why it matters (woven through — anchor here ~8:30–9:00)

Two ideas to land clearly (can be VO over B-roll or a short graphic insert):

1. Frame geometry = water flow. The downpipe needs fall — a consistent downhill gradient — or water pools instead of running. The frame is what sets and holds that fall. Rigidity isn't vanity; a frame that flexes loses its gradient under load, and a sagging downpipe is a stagnant downpipe.

2. Biofiltration is alive and slow. The media is just a home; the cleaning is done by a bacterial colony that takes weeks to establish. Starting it early ("cycling") means the biology is mature before it's needed — a standard move in [aquaponics / aquarium] keeping, applied here to a water-harvesting build.

On-screen graphic idea: simple side-view diagram — dome → downpipe on its fall → bio-filter → pond, with the filter labelled "cycling since Week 6."

Honest beat (~9:00–9:20)

To-camera (draft — pick whichever is true on the day):

[If something went sideways:] Real talk — [e.g. one footing came out [slightly off / not level] and I had to redo it, which ate an hour I didn't have. / the first bracing layout wasn't rigid enough and I had to add more than I'd planned.] Annoying, but better to find it now than when there's water in the line.
[If it went smoothly:] Honestly this one went better than I expected — but I'm a bit nervous about the bio-filter, because it's the one job today I can't check by looking at it. I won't really know if it's cycling properly for a couple of weeks. So that's the gamble of starting early — you commit and you wait.

Outro + next-week tease (~9:20–10:00)

Shot: Drone progress shot — pull up and back off the finished frame, dome in shot, the whole site in context. The frame reads as small but deliberate against the bigger build.

To-camera (draft):

So that's week six. The frame's up, it's plumb, it's braced, and it's not going anywhere. The bio-filter's quietly doing its thing in the corner. None of it moves water yet — but it's all ready for the thing that does.
Because next week, the downpipe itself goes on. All these mounting points I set today finally get something hung off them, and we find out if my fall is right. That's going to be a good one.
If you want to watch any of this happen in real time — and watch me check on that filter — the nightly livestream's on the other channel, [channel name / link]. Cheers for watching. See you next week.

On-screen text: NEXT WEEK → Installing the downpipe / Nightly livestream: [link] / @atmosfarm

Shot list

#	Shot	Type	Notes
1	Hand shoves finished upright, no movement	POV / tight	Cold open + callback in seg 3
2	Footing/fixing montage	B-roll, fast cuts	For cold open
3	Stephen intro to-camera, footprint behind	Handheld	Layout marks visible
4	String lines / marking paint, footprint	Overhead / high	Layout
5	Sight down the run, line straight	POV	Layout
6	Footings going in	Hands-in-frame	Confirm method
7	Raising first upright	POV	Temporary prop/clamp
8	Spirit level bubble settling	Macro	Money shot, hold it
9	Bare uprights standing, unbraced	Wide	Deliberately gawky
10	Fitting diagonal braces	POV / over-shoulder	The work
11	Geometry forming, rectangles → triangles	B-roll	Let shape tell story
12	Hand-load wreck test	2 angles	Callback to shot 1
13	Marking/fixing downpipe mounts to fall	Mid / detail	Geometry payoff
14	Filter media into vessel	Close-up	Confirm media type
15	Adding seed culture / starter	POV	Confirm method
16	Writing start-date label	Detail	Proof of head start
17	Drone pull-back off finished frame + dome	Drone	Progress shot, outro
18	Night b-roll under work lights [optional]	B-roll	If frame/filter runs late

On-screen text / graphics

WEEK 6 — DOWNPIPE FRAME (cold open)
TODAY: 1. Frame up + braced 2. Seed the bio-filter (early) (intro)
Footings + fixings — set before anything goes vertical
Plumb on two faces — every upright, no exceptions
Bracing = triangles = rigid
Mounts set to fall — water has to run downhill
Bio-filter seeded: [DATE] — cycling now, ready by Week 10
Bacteria don't hurry
Diagram: dome → downpipe (on fall) → bio-filter (cycling) → pond
NEXT WEEK → Installing the downpipe
Nightly livestream: [link] / @atmosfarm

Runtime breakdown

Section	Start	Duration
Cold open	0:00	0:30
Intro to-camera	0:30	1:00
Seg 1 — Layout, footings & fixings	1:30	2:00
Seg 2 — Uprights up & plumb	3:30	2:00
Seg 3 — Cross-bracing + mounting points	5:30	2:00
Seg 4 — Seeding the bio-filter	7:30	1:00
Science / why it matters	8:30	0:30
Honest beat	9:00	0:20
Outro + next-week tease	9:20	0:40
Total		10:00

PLACEHOLDER draft. Confirm bracketed assumptions on the day: footing/fixing method, brace count/config, downpipe fall spec, filter media type, seeding/cycling method, livestream channel link, prior-week callback.

↑ top

Atmosfarm — Weekly Build Video 07: Downpipe

Build week: 7 of 14 | Runtime: ~10 min | Status: PLACEHOLDER (draft VO/to-camera + shot direction — refine later. Assumptions in [brackets].)

Logline

Last week we built the frame to carry the water. This week we hang the pipe, chase the falls along the full run, seal every joint — and for the first time, send real water down the path to the pond.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): Push-fit and union joints, not glued solid — the downpipe breaks down into sections.

Cold open (0:00–0:30)

Visuals: Tight, wet, kinetic. Close on a downpipe outlet — a beat of nothing — then a first slug of water surges out and runs. Slow-mo on the splash hitting the collection point. Cut to drone rising fast over the dome and the new pipe run snaking off toward the pond.

VO (over the surge):

"All week it's been brackets, levels, and sealant. The whole point of it was this one moment — does the water actually go where I told it to?"

SFX: real water rush, no music yet — let the sound carry the open. Music stabs in on the drone reveal.

On-screen: ATMOSFARM / WEEK 7 — DOWNPIPE / Day [XX] of 100

Intro to-camera (0:30–~1:30)

Setting: Standing at the base of the frame, hand resting on a freshly mounted pipe section. Daylight. [Adelaide / build site]

To-camera (conversational):

"Right — week seven. Last week was the frame, the skeleton that holds all this off the ground. This week it gets its plumbing.
The job's actually pretty simple to say out loud: mount the downpipe sections onto that frame, set the fall — the gradient — so water runs downhill the whole way and never sits, seal the joints, leak-check them, and route the bottom end out to where the pond's going to be.
Simple to say. The thing that'll make or break it is the fall. Get the gradient wrong over a run this long and you get standing water, you get backflow, you get a slow drip out of a joint you swore was sealed. So today is slow, fussy, spirit-level work — and then, if it all holds, the first water test.
Let's hang some pipe."

Shot notes: Single clean take, slight handheld. Cut away mid-line to a pan along the empty frame so the viewer sees the run before we fill it.

Build segments

Segment 1 — Mounting the sections (1:30–3:30)

To-camera / VO mix:

"Pipe goes up in sections. I'm dry-fitting the whole run first — no sealant, just brackets loose — because once you commit to the fall you don't want to be unpicking it.
[Assumption: downpipe is [90mm PVC stormwater] running off the frame's high point down to the collection elbow.] Each section sits in a bracket, and the brackets are what set the height — so every bracket is a decision about the fall."

Shot notes:

Lifting a section into the first bracket, framed long so you feel the length of the run.
Insert close-up: bracket screw biting into the frame conduit. [Confirm fixing method — saddle clamp vs U-bolt — bracket-to-conduit detail.]
Time-lapse the dry-fit going up section by section.
Overhead/ladder shot looking down the line of the run.

Honest note to drop in: brief aside if a section won't sit right — "yeah, that one's fighting me" — keep it.

Segment 2 — Setting and checking the falls (3:30–5:30)

To-camera / VO:

"This is the part that matters. The fall. [Assumption: aiming for roughly a 1-in-100 gradient — about 10mm of drop per metre.] Too flat and water pools. Too steep and over a long run you run out of height before you reach the pond.
So I'm working bracket to bracket with the level, dropping each one a touch lower than the last, checking, adjusting, checking again. It's tedious and it's the whole ball game."

Shot notes:

Hero shot: spirit level laid along a section, bubble in frame, slight rack focus to the builder's eye.
Insert: tape measure / [laser level] showing drop between two points.
Hands easing a bracket down a few mm and re-checking.
Wide to show the gentle slope of the whole completed run against the horizon.

On-screen graphic cue: simple animated line showing the fall exaggerated, with ~10mm / metre label. [Confirm final gradient figure before animating.]

Segment 3 — Sealing, leak-check, routing to pond (5:30–7:30)

To-camera / VO:

"Falls are set, so now I lock it in. Every joint gets [solvent cement / rubber seal — confirm], and then I final-tighten the brackets and supports so nothing can shift and undo all that fall I just set.
Then the leak check. I'm running [a hose / a filled section] through and watching every joint for weeping before I trust it with the real thing.
And the last bit — routing the collection point out to the pond. This is where all of it ends up. Everything the dome captures funnels down to this one outlet and heads for [the pond location]."

Shot notes:

Close macro on a joint being sealed — sealant bead, the push-and-twist home.
Leak-check: water at a joint, hold on it, dry = good. If one weeps, FILM IT — that's honest gold.
Torque/tighten brackets, satisfying final cinching.
Walk-and-talk following the pipe from frame to the pond ground location, builder tracing the path with his hand.

The science / why it matters

VO over diagram/graphics:

"Here's why the fall obsession is the whole point. Atmosfarm pulls water out of the air — but capturing it is only half the trick. You then have to move it, cleanly, without losing it to evaporation, leaks, or stagnation.
Gravity is free. A correct fall means the water moves itself, every drop, every time, with no pump and no power on this leg of the system. Get the gradient right once and it works forever. Get it wrong and you're fighting standing water for the life of the build.
This downpipe is the artery. The core unit makes the water — this is how it gets to where it's stored."

On-screen: clean schematic — capture point → downpipe (with fall arrow) → collection elbow → pond. Label each stage. [Reuse the system overview graphic from the topic videos for consistency.]

Prototype test (the first water-flow / drainage test)

Setup to-camera:

"Okay. Moment of truth. First water test. I'm going to send water in at the top and see if it runs the full path, behaves on the falls, reaches the pond — and stays inside the pipe the whole way."

The test:

Pour/feed water in at the high point. [Source: hose / buckets — confirm.]
Follow it down. Hold on the outlet at the pond end for the payoff.
Watch the falls: is it running clean, or pooling anywhere?

Reaction (keep it real, whatever happens):

"[If it works:] There it is — full run, no pooling, straight to the pond. That's the artery open."
"[If it leaks/pools:] Right, see that — it's catching there. That's a fall I need to drop. This is exactly why you test before you walk away."

Shot notes:

Multi-angle: GoPro at the outlet, drone overhead catching the water's progress along the run, handheld for the reaction.
Slow-mo on first water out the pond end.
If there's a fault, document the fix in a short tag — don't hide it.

Honest beat

To-camera, quieter, end of day [or night-shoot — lighting is set up]:

"Not glamorous, this one. A week of bubbles in a spirit level and sealant on my hands. But this is the kind of work nobody sees and everything depends on. If the fall's wrong, none of the clever stuff upstream matters — the water just sits there.
[If test was clean:] First water down the path though. That felt good. [If it wasn't:] Got a couple of joints to chase tomorrow — that's the build."

Shot notes: Golden hour or night-light. Builder sitting on the frame base, pipe in shot. Let a silence sit.

Outro + next-week tease

To-camera:

"So that's the frame plumbed and the path proven. Water can get from the dome to the pond now.
Next week — the big one. The water-2-air core unit. The heart of the whole build. The thing that actually pulls water out of thin air. Everything we've built so far has been getting ready for this.
If you want to watch it happen as it happens, the nightly livestream's on [the other channel] every evening. I'll see you there. Cheers."

Shot notes: Drone pull-back off the dome and pipe run into the [Adelaide] landscape. End card.

On-screen: NEXT WEEK — THE CORE UNIT / Water-2-Air + subscribe / livestream channel handle.

Shot list

Cold open: water surge out of outlet (slow-mo, GoPro)
Cold open: drone rise over dome + pipe run
Intro to-camera at frame base (daylight)
Pan along empty frame (the run before water)
Section lifted into first bracket (long)
Macro: bracket fixing biting into conduit
Time-lapse: dry-fit run going up
Overhead looking down the run line
Hero: spirit level + bubble, rack focus
Insert: [laser level] / tape showing drop
Hands lowering a bracket a few mm
Wide: finished slope against horizon
Macro: joint sealing / push-home
Leak-check held on a joint (dry = pass; weep = film it)
Final bracket tightening
Walk-and-talk: pipe traced frame → pond, hand on pipe
Test: water in at top
Test: GoPro at pond outlet, slow-mo first water
Test: drone overhead tracking water down run
Test: live reaction
Honest beat: builder on frame base, golden hour / night-light
Outro: drone pull-back into landscape
End card

On-screen text / graphics

ATMOSFARM / WEEK 7 — DOWNPIPE / Day [XX] of 100 (cold open)
Fall animation: exaggerated slope line, ~10mm / metre label [confirm figure]
System schematic: capture → downpipe (fall arrow) → collection → pond
Lower-thirds: [material/spec callouts — pipe diameter, bracket type] (confirm specs)
NEXT WEEK — THE CORE UNIT / Water-2-Air
Subscribe + nightly livestream handle [other channel]

Runtime breakdown

Segment	Start	Duration
Cold open	0:00	0:30
Intro to-camera	0:30	1:00
Seg 1 — Mounting sections	1:30	2:00
Seg 2 — Setting the falls	3:30	2:00
Seg 3 — Seal / leak-check / route	5:30	2:00
Science / why it matters	7:30	0:45
Prototype test (water flow)	8:15	1:00
Honest beat	9:15	0:25
Outro + tease	9:40	0:20
Total		~10:00

↑ top

Atmosfarm — Weekly Build Video 08: Water-2-Air

Build Week: 8 of 14 | Target runtime: ~10:00 | Status: PLACEHOLDER (draft — refine later) | FLAGSHIP / HERO EPISODE

Logline

This is the one. The water-2-air unit is the heart of the whole project — the piece that actually turns moving air into liquid water. This week I assemble it, tie it into the intake on the air side and the downpipe on the water side, seal the condensation surfaces, fit the Venturi, and run the full path end-to-end. And then, on camera, I watch the first droplets form. No tricks, no overclaiming — just the real physics of pulling water out of the sky.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): The core unit is a self-contained module — disconnect two unions and it's out.

Cold open (0:00–0:30)

Visual: MACRO close-up, locked off, shallow focus — a cold metal surface, dark background. Nothing happening. Hold. Then, faint at first, a bead of moisture swells out of the surface, catches the night-work light, and runs. Pull this from the hero test footage later in the episode — open on the payoff.

VO (over the macro, quiet, almost reverent — but grounded):

"That's it. That right there — that's the whole reason I'm building this thing."
"Water. Out of the air. No tank, no tap. Just the right surface, at the right temperature, and a bit of physics doing the work."

VO (beat, honest):

"It's not free. It's not magic. But it's real — and this week I built the part that makes it happen."

Hard cut to title card: WATER-2-AIR.

Intro to-camera (0:30–~1:30)

Visual: Stephen to-camera, standing beside the partly-assembled water-2-air unit / the dome behind him. Handheld, daylight. Energy is up — he knows this is the big one.

To-camera (conversational, raises the stakes honestly):

"Welcome back to Atmosfarm. If you're new — I'm Stephen, I'm building a geodesic dome out here in [South Australia] that harvests water out of the air, and we're at week eight of fourteen."
"I'm not going to undersell it: this is the most important week of the entire build. Everything before this — the foundation, the frame, the intake ducting — all of it has been leading to this one component. The water-2-air unit. It's the part that actually condenses moisture out of the air and sends it down to the pond. It's the engine."
"So this week I'm assembling it, bolting it into the air intake on one side, plumbing it into the downpipe on the other, sealing up the condensation surfaces, fitting the Venturi, and then — the bit I've been waiting a hundred days for — I run a real condensation test. On camera. And we find out if any of this works."

On-screen text: Day [50] / 100 and Week 8 — Water-2-Air and THE ENGINE

Build segments

Segment 1 — Assembling & positioning the unit (1:30–3:30)

To-camera / VO:

"First, get it built and get it in place. The water-2-air unit sits [assumption: at the high point of the intake path, where the air's already been funneled and sped up]. It has to line up with two things at once — the intake duct coming in on the air side, and the downpipe heading down to the collection pond on the water side. If it's even a bit off, I'm fighting it on both connections."
"So I'm dry-fitting the whole thing first. No sealant, no permanent fixings — just checking it lands where it's supposed to before I commit."

Shot notes:

POV of laying out the unit's components — get the weight and scale of it.
Close B-roll: hands aligning the unit to the intake flange / mount points.
Wide: the unit lifted/positioned into place [assumption: solo with a strap or brace, or a second pair of hands — flag if so].
Timelapse of the dry-fit and adjustment.

On-screen text: Air side ← unit → water side and Dry-fit FIRST, commit later

Segment 2 — Connecting air side (intake) & water side (downpipe) (3:30–5:30)

To-camera / VO:

"Two connections, two completely different jobs. On the air side, I'm joining the unit to the intake duct — and that has to be airtight, because any air that leaks out around the join is air that never touched the cold surface. Leaks here are lost water, plain and simple."
"On the water side, it's the opposite problem. Gravity does the work — the condensate just has to be able to run downhill, cleanly, into the downpipe and down to the pond. So I want fall on that pipe the whole way, no traps, no flat spots where water pools and sits."
"[assumption: I'm using (gasket / foam tape / sealant) on the air joins and (rubber boot / solvent-weld fitting) on the water side.]"

Shot notes:

Close-up: air-side join going together — gasket/seal seating.
Close-up: water-side downpipe connection, with a level checking the fall.
B-roll: a hand following the downpipe run, showing it drop continuously.
Quick graphic overlay: air path (blue arrows in) vs water path (droplet arrows down).

On-screen text: Air joins = airtight and Water joins = downhill, no traps

Segment 3 — Condensation surfaces & the Venturi (5:30–7:00)

To-camera / VO:

"Now the two pieces that actually make it work. First, the condensation surfaces — these are the cold faces the moist air runs across, where the water gives up and turns liquid. I'm fitting them so the air is forced to flow right over them, not around them, and sealing the edges so nothing sneaks past."
"And second — the Venturi. This is a deliberate narrowing in the air path. When you squeeze a moving airflow through a tighter gap, it speeds up and the pressure drops. That faster, lower-pressure air across the cold surface is what helps the condensation along — and it's a big part of how this rig gets more water for less energy. We'll get into exactly why in a second."

Shot notes:

MACRO B-roll: the condensation surface up close — texture, finish [assumption: (finned / smooth metal) surface].
Close-up: fitting and sealing the surfaces into the flow path.
Close-up: the Venturi narrowing — show the constriction clearly.
Optional simple animated overlay: airflow squeezing through the Venturi, speeding up.

On-screen text: Condensation surface — where air gives up its water and Venturi — squeeze the airflow, speed it up

Segment 4 — Verifying the full path end-to-end (7:00–7:40)

To-camera / VO:

"Before any test, I walk the whole thing. Intake → unit → downpipe → pond. Every join. I want to confirm air can get all the way through on one side, and water has a clear run all the way down on the other. If there's a blockage or a leak anywhere in that chain, the test is meaningless."

Shot notes:

POV travelling shot following the full path: intake, into the unit, out the downpipe.
Close-up: smoke/tissue check at the joins for air leaks [assumption: smoke pencil or incense stick].
Pour-test down the downpipe to confirm clean fall to the pond.

On-screen text: Intake → Unit → Downpipe → Pond ✓

The science / why it matters (7:40–8:40) — CENTREPIECE

Visual: To-camera, calm and clear, with a clean diagram overlay of the full system (intake funnel → Venturi → cold condensation surface → downpipe → pond). This is the most important 60 seconds in the whole series — slow down and get it right.

To-camera (the honest core of the project):

"Okay. Here's how this actually makes water — and I'm going to be straight with you about the physics, because the honesty is the whole point of this channel."
"Air always has some water in it as invisible vapour. Warm air holds more, cold air holds less. So if you take moist air and cool it down past its dew point, the air can't hold all that vapour anymore — and it condenses out as liquid water on the nearest cold surface. That's it. That's the entire trick. It's the same reason a cold drink sweats on a hot day."
"So the dome funnels air in, the Venturi speeds that air up and drops its pressure, and it gets driven across the cold condensation surfaces — and water beads out and runs down to the pond."

To-camera (the honest limit — say it plainly):

"Now — here's the part most 'water from air' videos won't tell you. This is energy-limited. The amount of water you can pull out depends on how humid the air is and how much energy you've got to move it and cool that surface. You can't beat the physics. Anyone promising unlimited free water out of thin air is selling you something."
"What the funnel and the Venturi and the turbine actually do is make it more efficient — they get more water per unit of energy. Realistically we're talking something like [assumption: 20 to 40%] better efficiency than a dumb fan-and-coil setup. That's not 'free water.' That's smarter water. And honestly, that's the part I'm proud of."

On-screen text / graphics:

Diagram: warm moist air → cooled past dew point → liquid water
Dew point: cool air enough and it can't hold its water
This is energy-limited — NOT free water
Funnel + Venturi + turbine = ~[20–40%] more efficient, not magic

Prototype test (8:40–9:20) — HERO FOOTAGE

Visual: The condensation + Venturi test. This is the most important footage on the channel — shoot it like it. Run it [assumption: at night, when air is coolest and the surface runs coldest — night-work lighting is set up]. Get airflow moving through the unit and watch the surface.

To-camera (live, real, unscripted — let the nerves show):

"Alright. This is the moment. Airflow's on, surface is cold, the path's sealed. Let's see if it actually does it."

Then GO TO MACRO. Let the footage carry it — minimal talking.

MACRO B-roll (the hero takes — flag every one of these for the launch edit):

Locked-off macro on the condensation surface. Hold long. Watch the first beads appear.
Beads growing, merging, and the first drop running down the surface.
Drop reaching the bottom edge and falling — track it if possible.
Wider reset: water visibly collecting and moving toward the downpipe.

To-camera (after, genuine reaction — whatever it really is):

"[assumption: There it is. That's water. That's actual water, out of the actual air. I've been picturing this moment for months and it's better than I thought.]"
"[OR, if it underperforms — keep it honest:] Okay — it's working, but slower than I hoped. Less than I wanted. But it IS condensing, and now I know exactly what I'm tuning for the rest of the build."

On-screen text: FIRST WATER and Macro — real droplets, no VFX

[PRODUCTION NOTE: Flag ALL macro condensation takes as HERO for the launch edit / flagship film / trailer. This is the single most important capture day of the project. Shoot generous coverage, multiple angles, multiple takes. Back up the cards immediately.]

Honest beat (9:20–9:40)

Visual: To-camera, end of the night, tired, lit by the night-work rig. Quiet.

To-camera (real):

"Honestly? [assumption: It works — but a few drops on a test surface is a long way from filling a pond. The real numbers, the litres-per-day, the days it's too dry to pull anything — all of that is still ahead of me.] But for the first time in a hundred days, I've seen it do the one thing the entire project is named after. The engine runs. Everything from here is making it better."

Outro + next-week tease (9:40–10:00)

Visual: Drone progress shot — pull up and away from the dome at [dawn / dusk], the water-2-air unit now part of the structure. The thing is starting to look like a working machine.

To-camera / VO:

"So that's week eight — the biggest one. The water-2-air unit's built, it's tied into the intake and the downpipe, it's sealed, and it makes water. The heart of Atmosfarm is beating."
"But water out of the air isn't drinking water yet — not until it's clean. So next week I'm building the filtration system: getting this from 'condensate' to something you'd actually want to drink."
"The nightly livestream's on every night on the [second channel] if you want the raw version. Subscribe so you don't miss the filtration build — and I'll see you next week."

On-screen text: NEXT WEEK → Filtration: from condensate to drinkable + Subscribe + nightly livestream @atmosfarm

Shot list (quick reference)

MACRO hero — first droplet forming (cold open AND test) — HERO, flag for launch edit
Title card — WATER-2-AIR
To-camera intro — stakes, beside the unit
POV — laying out / handling unit components
Close-up — aligning unit to intake mount
Wide — unit positioned into place
Timelapse — dry-fit & adjustment
Close-up — air-side join / gasket seating
Close-up — water-side downpipe join with level (checking fall)
B-roll — hand following downpipe run downhill
Graphic overlay — air path vs water path
MACRO B-roll — condensation surface texture/finish
Close-up — fitting & sealing condensation surfaces
Close-up — the Venturi narrowing
Animated overlay — airflow speeding through Venturi
POV travelling — full path intake → unit → downpipe → pond
Close-up — smoke/leak check at joins
Pour-test down downpipe
To-camera — science centrepiece (with system diagram overlay)
HERO TEST — locked-off macro: first beads, growth, merge, run, drop falling
HERO TEST — wider reset: water collecting toward downpipe
To-camera — live reaction to the test
To-camera — honest end-of-night beat (night-work light)
Drone progress pull-up/away — dome with unit installed

On-screen text / graphics

Day [50] / 100
Week 8 — Water-2-Air
THE ENGINE
Air side ← unit → water side
Dry-fit FIRST, commit later
Air joins = airtight / Water joins = downhill, no traps
Condensation surface — where air gives up its water
Venturi — squeeze the airflow, speed it up
Intake → Unit → Downpipe → Pond ✓
Diagram: warm moist air → cooled past dew point → liquid water
Dew point: cool air enough and it can't hold its water
This is energy-limited — NOT free water
Funnel + Venturi + turbine = ~[20–40%] more efficient, not magic
FIRST WATER
Macro — real droplets, no VFX
NEXT WEEK → Filtration: from condensate to drinkable
Subscribe + nightly livestream @atmosfarm
Lower-third (first appearance): Stephen — Atmosfarm

Runtime breakdown

Section	Start	Length
Cold open (macro payoff)	0:00	0:30
Intro to-camera	0:30	1:00
Seg 1 — Assemble & position unit	1:30	2:00
Seg 2 — Connect air side & water side	3:30	2:00
Seg 3 — Condensation surfaces & Venturi	5:30	1:30
Seg 4 — Verify full path end-to-end	7:00	0:40
Science / why it matters (centrepiece)	7:40	1:00
Prototype test (hero macro)	8:40	0:40
Honest beat	9:20	0:20
Outro + next-week tease	9:40	0:20
Total		~10:00

↑ top

Atmosfarm — Weekly Build Video 09: Filtration System

Build week: Week 9 of [14] · Runtime: ~10 min · Status: PLACEHOLDER (draft VO/to-camera + shot direction — refine before shoot)

Logline

Before a single drop circulates, the water needs somewhere clean to come back to. This week I build the two filters every healthy pond runs on — one mechanical, one biological — plumb them together, leak-test the whole rig under flow, and run a before/after clarity test that's honestly the most satisfying thing I've filmed yet.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): Skid-mounted filter rig — wheel it out in one piece.

Cold open (0:00–0:30)

Visual: Tight macro shot — cloudy, sediment-heavy [tank/bucket] water swirling. Hold. Then hard cut to the same water pouring out the far end of the rig: clear. Pull focus to my hands holding both glass jars side by side against the morning light.

VO (over the cut):

"This is the water going in. [beat] And this is the same water, thirty seconds later, after it's been through the thing I built this week. No chemicals. Just gravel, sponge, and a few billion bacteria doing me a favour."

On-screen text: WEEK 09 — FILTRATION (lower third, fades)

Shot note: Shoot the jars handheld, slightly loose. Don't over-light — let it look real. This is the hook; the clarity reveal is the whole episode in two seconds.

Intro to-camera (0:30–~1:30)

Setting: [Outside by the dome / next to the filter bench]. Daytime. Conversational, holding a length of [pipe] or the bio-media.

To-camera (draft):

"Alright — week nine. Last few weeks have been a lot of metal and structure, the dome's actually starting to look like a dome. But this week we go invisible. Plumbing. Filtration.
Here's the thing nobody tells you about water harvesting: collecting the water is only half the job. If you want it to stay good — if you want a pond, fish eventually, plants, the whole aquaponics side — the water has to keep moving and it has to keep getting cleaned. Forever. Quietly. In the background.
So before the pond goes live next week, I'm building the filters first. On purpose. I want the water to have somewhere clean to circulate from literally day one, instead of going green and me scrambling to fix it after.
Two filters. A mechanical one and a biological one. By the end of today they're plumbed together, leak-tested, and I'll show you the clarity test — which, no spoilers, but the cold open kind of gave it away."

Shot note: One or two takes, keep the stumble if it's charming. Cut to a quick 3-shot montage of today's goal: mechanical filter → bio housing → jars. Teases structure of the episode.

Build segments

Segment 1 — Mechanical filter stage (1:30–4:00)

To-camera / VO (draft):

"First stage: mechanical. This is the dumb-but-essential one. Its whole job is to physically catch the big stuff — sediment, leaf bits, gunk — before it ever reaches the biology. Think of it as the bouncer.
I'm running [a settlement chamber into a graded sponge stack — coarse foam first, then medium, then fine]. Water hits the coarse layer, drops its heaviest load, and gets progressively polished as it works through. The trick is the order: coarse to fine. Get it backwards and the fine pad clogs in a day and you're elbow-deep in it every weekend."

Build beats to film:

Cutting / fitting the [filter housing] to size on the bench.
Laying in the foam stack, coarse → fine. Hold a shot of each layer.
First fittings going on — [barbed connectors / pipe glue]. Honest moment if a fitting doesn't seat.

VO over assembly:

"I'm oversizing this stage a bit on purpose. A filter that's slightly too big just means I clean it less often. A filter that's too small means it's the bottleneck for the entire system. Bigger is lazier, and lazier is the goal."

Shot notes: Overhead bench shot for the layer-stacking (graphic-friendly). Close macro on water-path direction. [Assumption: housing is a repurposed [drum/tote] — confirm and reshoot intro line if it's a commercial canister.]

Segment 2 — Bio-filter housing: install + connect (4:00–6:30)

To-camera (draft):

"Stage two is where it gets cool. The biological filter. Back in week six I seeded the media for this — remember the [media] I dropped into the [tank] and basically left to marinate? That's been quietly growing a colony of bacteria this whole time, and those are the actual workhorses.
The mechanical filter catches what you can see. The bio filter handles what you can't — the dissolved stuff, ammonia from anything living in the water, breaking it down into things plants can actually use. You genuinely need both. One's a sieve, one's a tiny invisible sewage-treatment city."

Build beats to film:

Lifting the seeded media across from its week-6 home — get a look at it, comment on whether it smells/looks colonised.
Setting the bio housing in position, connecting it downstream of the mechanical stage.
Plumbing the two stages together — the satisfying "click" of the loop taking shape.

VO (draft):

"Order matters here too — mechanical then biological, always. The bacteria are delicate. If I let raw sediment smother them, the colony suffocates and I lose six weeks of waiting in an afternoon. So everything dirty gets caught up front, and the bio stage only ever sees pre-cleaned water."

Shot notes: Hero shot of the two filters sitting side by side, plumbed. Drone/high angle optional here to show the rig in context with the dome behind. [Confirm bio-media type — moving-bed / static / lava rock — line above assumes a static colonised media.]

Segment 3 — Leak-test under flow + stub the loop (6:30–8:00)

To-camera (draft):

"Right — moment of truth. Everything's connected, but connected isn't the same as watertight. I'm putting it under flow and walking every joint."

Build beats to film:

Priming the [pump], water moving through the full rig for the first time.
Slow walk along the plumbing, hand on each joint, looking for weeping.
Find at least one drip (there's always one). Fix it on camera — [re-seat / extra clamp / [thread tape]].
Stubbing the inlet and outlet: capping off the points where the pond loop connects next week.

VO:

"Found one — of course I did. [point] Always the joint you were most confident about. [fix it] ... and now it's dry. I'm leaving the inlet and outlet stubbed and capped right here, so next week the pond just plugs straight in. Filtration's done and waiting before there's anything to filter."

Shot notes: Close macro on the drip vs. the fixed-and-dry joint — before/after pair, same framing. Keep the fix unglamorous and real.

The science / why it matters

Format: To-camera, possibly at [whiteboard / with the rig as the prop]. Can carry simple motion graphics.

To-camera (draft):

"Quick why-it-matters, because this is the part that decides whether this whole project lives or dies long-term.
Mechanical filtration removes what you can see — particles, sediment, debris. It's purely physical: stuff too big to pass gets trapped. Easy.
Biological filtration removes what you can't see, and it's the one people skip and then wonder why their water turns into a swamp. Living water produces ammonia — toxic. Bacteria on the bio-media eat ammonia and turn it into nitrite, then a second colony turns nitrite into nitrate, which is basically plant food. That whole chain is the nitrogen cycle, and it's what makes aquaponics even possible.
You need both because they do completely different jobs. Skip mechanical and your bio filter clogs and dies. Skip biological and the water's clear but slowly poisoning anything in it. Together, they're a closed, self-cleaning loop. That's the entire point of building an Atmosfarm — water that looks after itself."

On-screen graphic: Simple flow diagram — RAW WATER → [Mechanical: traps particles] → [Biological: ammonia → nitrite → nitrate] → CLEAN WATER → back to pond. Keep it clean, hand-drawn-ish to match the journal tone.

Prototype test — Filter throughput / water-clarity test (8:00–8:45)

To-camera setup (draft):

"Okay — the test I've been wanting to run all week. I've got a batch of deliberately filthy water here — [sediment / [soil] stirred in] — and I want to see two things: how fast the rig can push it through, and how clear it comes out the other side."

Test beats to film:

Fill jar A from the dirty input. Set it aside as the control.
Run the batch through the full rig. [Time a measured volume — e.g. [X] litres — for a rough throughput number.]
Fill jar B from the clean output.
Set jar A and jar B side by side, backlit. The reveal.

VO / to-camera reaction:

"[hold the jars up] ... that's [X] litres in [Y] seconds, and look at that. Same water. I'm not going to pretend that's not deeply satisfying. [Assumption: clarity visibly improves on first pass; if not, note it honestly and explain the bio side needs more cycling time.]"

Shot notes: Backlit jar comparison is the money shot — match the cold open framing so it pays off. Optional macro of clear water trickling from the outlet. Lower-third with the throughput number as on-screen text.

Honest beat (8:45–9:15)

To-camera (draft) — unpolished, real:

"Couple of honest things. One — the clarity test only proves the mechanical side really. The biological filter? I can't show you that working. There's no satisfying before/after for bacteria. I have to trust that the colony I seeded in week six is actually colonising, and the real proof won't come for weeks, once there's a living load on the system. So there's a bit of faith baked into today.
Two — [a fitting / the flow rate / a leak I had to chase] cost me a good [chunk of the afternoon / a re-do], and I'm not 100% sure the throughput's where I want it for a full pond yet. Might have to revisit. That's the build. Nothing's ever quite finished, it just gets to 'good enough to move forward.'"

Shot note: Golden hour or [night-work lighting] if it ran late — lean into the long-day feel. Single take, let it breathe.

Outro + next-week tease (9:15–~10:00)

To-camera (draft):

"So that's filtration. Mechanical built and plumbed, bio housing in and connected, whole rig leak-tested under flow, and the inlet and outlet stubbed and waiting. The cleaning system is alive before there's anything to clean — exactly how I wanted it.
And next week is the big one. The pond goes in. The whole water circuit goes live — pond, pump, these filters, the start of the aquaponics. Everything I've been building toward in the background finally connects into one loop. Honestly can't wait.
If you want the unedited version, I'm livestreaming the build for an hour every night on [the second channel] — link below. Cheers for watching. See you next week."

Shot notes: End on a drone progress shot — pull up and away from the dome with the filter rig visible, revealing the site and [the Adelaide / SA] landscape. Hold for the outro card.

Outro card: NEXT WEEK — THE POND GOES LIVE + subscribe / livestream channel prompt.

Shot list

Cold open: macro dirty water swirling (tank/bucket)
Cold open: clear water pouring from outlet
Cold open: hands holding both jars, backlit, morning light
Intro to-camera by the dome / filter bench
3-shot goal montage (mechanical / bio / jars)
Mechanical: housing cut & fit on bench
Mechanical: foam stack laid coarse → fine (overhead)
Mechanical: fittings going on (+ honest mis-seat moment)
Bio: lifting week-6 seeded media across, inspection close-up
Bio: housing set in position & connected
Bio: two filters plumbed side by side (hero shot)
Leak test: pump primed, first flow through rig
Leak test: walking the joints, finding the drip (macro)
Leak test: drip fixed, joint dry (matching macro)
Stubbing/capping inlet & outlet
Science to-camera (whiteboard / rig as prop)
Clarity test: jar A dirty control
Clarity test: batch running through, throughput timing
Clarity test: jar B clean output
Clarity test: A vs B backlit reveal (match cold open)
Honest beat to-camera (golden hour / night lighting)
Outro to-camera
Drone progress shot — pull up/away from dome + rig + landscape
B-roll buffer: water trickling from outlet, hands on plumbing, media texture

On-screen text / graphics

WEEK 09 — FILTRATION (cold open lower third)
Nitrogen-cycle / two-stage flow diagram: RAW WATER → MECHANICAL → BIOLOGICAL → CLEAN WATER → POND
Layer-order callout on mechanical stage: COARSE → MEDIUM → FINE
Throughput lower third: [X] L in [Y] sec during clarity test
Jar comparison labels: IN / OUT
Outro card: NEXT WEEK — THE POND GOES LIVE
Persistent: livestream channel handle / subscribe prompt (outro)

Runtime breakdown

Section	Start	Duration
Cold open	0:00	0:30
Intro to-camera	0:30	1:00
Segment 1 — Mechanical filter	1:30	2:30
Segment 2 — Bio-filter install	4:00	2:30
Segment 3 — Leak test + stub loop	6:30	1:30
The science / why it matters	8:00	— (folded into flow; see note)
Prototype test — clarity	8:00	0:45
Honest beat	8:45	0:30
Outro + next-week tease	9:15	0:45
Total		~10:00

Note: "The science" can either sit as its own ~0:45 block (trim Segment 1 or 3 by 0:45 to fit) or be woven into Segment 2's bio-filter explanation. Draft assumes it's largely folded in; pull it out as a standalone beat if the standalone graphic plays better. Reconcile exact timings in edit.

↑ top

Atmosfarm — Weekly Build Video 10: Pond & Aquaponics

Build week: Week 10 of [14] · Runtime: ~10 min · Status: PLACEHOLDER (draft VO/to-camera + shot direction — refine later. Assumptions marked in [brackets].)

Logline

The pond is a portable tank — we position it, level it, fill it and plumb it into a living loop — tank to filter to grow-beds and back — and run the whole water circuit end to end for the first time. We do it now, while the dome is still open and access is easy, because the tank carries its own lid and doesn't wait on the skin.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): The pond is a portable TANK, not a dug hole — drain it, and it moves with everything else.

Cold open (0:00–0:30)

Visual: Tight, low shots of water moving — first trickle out of the return pipe into the grow-bed, then a wider reveal of the loop running. Quick cuts. No talking yet, just the sound of water and the pump hum. Hold on a final wide of the pond full and dark and still, dome ribs reflected on the surface.

VO (over the last shot, quiet):

"Ten weeks in. For the first time, the water's actually going somewhere. Pond... filter... grow-beds... back to the pond. Round and round. Let's see if it holds."

On-screen text: WEEK 10 — POND & AQUAPONICS (lower third, fades)

Intro to-camera (0:30–~1:30)

Setting: Stephen standing in the empty/just-dug pond hole, or on the edge of it, dome open and skeletal behind him. Daylight.

To-camera (conversational, not scripted tight):

"Alright — welcome back. Big week. We're doing the pond AND the aquaponics in one go, which sounds mad, but there's a reason for the timing.
Normally you'd think — surely you wait until the dome's closed in before you put a body of water in the middle of it? But here's the thing: the pond gets its own cover. Its own lid, basically. So it's not relying on the dome's outer skin at all. Which means I can do all the heavy water work NOW — getting the tank in and plumbed while the dome is wide open and access is easy. Once that skin goes on, access gets a lot tighter.
So this week: position the tank, level it, fill it, plumb the whole loop, and — fingers crossed — run the first full water circuit. Let's get into it."

Shot notes: One clean take ideally; cut in B-roll of the open dome and the tank going in as he talks.

Build segments

Segment 1 — Position & set the tank (1:30–3:30)

VO / to-camera (mix):

"First job — get the tank in and dead level. A pond that keeps fish and feeds plants needs a deep, cool, stable body of water where the fish ride out a hot Adelaide day, plus shallow planting zones up top. A poly tank / IBC setup gives us that without digging."

Shot notes:

Position and level the tank on its compacted base. Time-lapse the set-up.
Point out the depth, the fill line and the overflow on the tank — show it with your hands.
Marker pegs / string lines showing depth zones. Graphic overlay later (see graphics list).
Honest aside if the soil fights back — rocks, clay, water table. Leave it in.

Segment 2 — Plumb & fill (3:30–5:30)

VO / to-camera:

"Once it's level, fit the bulkhead and overflow and connect the loop. This is the bit you cannot rush — every fitting sealed, the overflow set to the right height. Then we start filling."

Shot notes:

Fitting the bulkhead/overflow and loop connections; checking for leaks as it fills.
TIME-LAPSE of the tank filling — set a locked-off camera, let it run. This is a hero shot. Water rising to the line, the loop coming alive.
Cut-in: the overflow point. "And here's where it gets locked in —" (lead into the overflow beat).

Overflow / intake lock-in (key build fact):

"This overflow height is not random. It has to line up with the downpipe and the intake plan — where the harvested water comes IN and where the excess goes OUT. So I'm setting the overflow level against that plan right now, before I commit. Get this wrong and the whole rainwater side fights the pond."

Shot notes: Show the overflow fitting, a level/laser against the downpipe reference. [Confirm exact overflow height number — TBD against intake plan.]

Segment 3 — Plumb the loop: pond → filter → grow-beds (5:30–7:00)

VO / to-camera, walking the loop:

"Right, this is the circuit. Follow the water with me. Pump pulls from the pond, here. Pushes it up through the filter — solids out, biology stays. Out of the filter, into the grow-beds — that's where the plants do their thing, they strip the nutrients out. And then clean water drains back down to the pond. That's the loop. Pond, filter, beds, pond."

Shot notes:

Continuous-ish walking shot following the water path start to finish — this is the spine of the segment. May need 2–3 takes stitched.
Fit the pump (in the pond / pump chamber [confirm location]).
Close-ups on each junction: pump outlet, filter in/out, grow-bed feed, return drain.
Bell siphon / flood-and-drain mechanism if used in the beds [confirm grow-bed style — media bed flood/drain vs. constant flow].

The science / why it matters (7:00–7:45)

VO over simple animated diagram (see graphics):

"Here's why this works. The fish produce waste — ammonia. On its own that's toxic. But bacteria in the filter and the grow-bed media turn that ammonia into nitrate — and nitrate is plant food. So the plants clean the water for the fish, the fish feed the plants, and you barely change the water at all. That's aquaponics. A nearly-closed loop.
And it sits right in the middle of an atmospheric water dome — water we're pulling out of the air, feeding a system that grows food. That's the whole point of this place."

Shot notes: Animated nitrogen-cycle loop graphic. Keep it clean and quick — don't lecture. [Confirm fish species and plant choices for accuracy — leave generic for now.]

Prototype test — FIRST FULL WATER-CIRCUIT TEST (7:45–8:45)

This is the week's big test. Film it properly.

To-camera, before:

"Okay. Everything's connected. Pond's full, pump's in, filter's primed, beds are ready. This is the first time the entire loop runs end to end. No idea if it all holds. Here goes."

Shot notes:

Hand on the switch. Real reaction — keep the camera rolling through the nervous bit.
Cover the full path on camera as it comes alive: pump kicks, water hits the filter, fills the beds, returns to the pond.
Watch for the failures and KEEP THEM: leaks at joints, flow too weak/too strong, beds flooding wrong, overflow behaving unexpectedly. Whatever happens, it's the honest record.
Wide shot of the whole loop running at once if the geometry allows.
[If a livestream ran the same night, note it: "the livestream crew watched this live on the other channel" — tie in the nightly 1-hr stream.]

To-camera, after (whatever the result):

"[If it works] ...it's running. The whole thing's running." / "[If it doesn't] ...okay, so we've got a [leak/flow] problem. That's tonight's job."

Honest beat (8:45–9:15)

To-camera, lower energy, real:

"Some honesty. [Placeholder — pick what actually happened: the tank was a pain to get dead level / the overflow level was a nightmare to lock against the downpipe / the first run sprung a leak at the filter / the pump flow wasn't what the beds wanted.] This is the bit nobody films usually. But that's the channel — it's a build journal, not a brochure. We problem-solve it and we keep going."

Shot notes: Late light or night-work lighting — we have night rigs set up, so a genuine after-dark fix-it shot fits here and looks great. [Confirm which problem to feature once shot.]

Outro + next-week tease (9:15–10:00)

To-camera, by the full pond, dome open above:

"So — the pond's in, it's full, and the water's moving in a loop for the first time. And remember, this whole thing's got its own cover, so it's safe and self-contained even with the dome wide open above it.
Which means next week... we close it in. The skin goes on the dome. After ten weeks of looking up through bare ribs at the sky — we finally put the cover on. You won't want to miss that one."

Shot notes:

Drone progress shot — pull up and out: pond full and dark in the centre, the open dome skeleton around it, the whole site in context. Strong closer.
Tease cut: 1–2 fast shots of skin/cover material on a pallet or a rib being prepped. [Confirm skin material/footage availability.]
End card / subscribe + nightly livestream plug.

Shot list

Cold open: tight water-moving macro shots (return into grow-bed, pump hum) + wide of full still pond with dome reflected
Intro to-camera in/at the pond hole, open dome behind
Tank position & level — time-lapse
At the tank pointing out depth, fill line + overflow
Depth-zone marker pegs / string lines
Bulkhead/overflow fit + loop connections
Pond-filling time-lapse (locked-off, hero shot)
Overflow fitting + level/laser against downpipe reference
Walking-the-loop continuous shot pond → filter → beds → pond
Pump fitting + each plumbing junction close-up
Grow-bed flood/drain mechanism [if used]
Science diagram screen-capture/animation
First full circuit test — switch-on, reaction, water travelling the whole loop
Honest-beat to-camera (night-work lighting)
Drone pull-up — full pond inside open dome
Next-week tease: skin/cover material B-roll

On-screen text / graphics

WEEK 10 — POND & AQUAPONICS opening lower third
Tank callout graphic: depth, fill line + overflow labelled
Overflow height callout — OVERFLOW LOCKED TO INTAKE PLAN [+ figure TBD]
Loop diagram labels as he walks: 1 PUMP → 2 FILTER → 3 GROW-BEDS → 4 RETURN
Nitrogen-cycle animation: FISH → AMMONIA → BACTERIA → NITRATE → PLANTS → CLEAN WATER
FIRST FULL CIRCUIT TEST title card before the prototype test
Day counter (consistent with series) [confirm format, e.g. DAY 6_/100]
Next-week teaser card: NEXT: SKINNING THE DOME
Persistent: subscribe + "nightly livestream on [stream channel name]"

Runtime breakdown

Section	Start	Length
Cold open	0:00	0:30
Intro to-camera	0:30	1:00
Seg 1 — Position & set the tank	1:30	2:00
Seg 2 — Plumb, fill, overflow	3:30	2:00
Seg 3 — Plumb the loop	5:30	1:30
The science / why it matters	7:00	0:45
Prototype test (first full circuit)	7:45	1:00
Honest beat	8:45	0:30
Outro + next-week tease	9:15	0:45
Total		10:00

↑ top

Atmosfarm — Weekly Build Video 11: Cover

Build week: 11 of [14] · Runtime: ~10 min · Status: PLACEHOLDER (draft VO / to-camera + shot direction — refine later; assumptions in [brackets])

Logline

The dome stops being a skeleton. This week we skin it — pre-cutting and numbering every triangle, working in rings from the base up, sealing the seams, and framing out the entrance so next week's forklift-wide doorway has somewhere to go. The day it finally looks finished.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): Every panel is numbered and clipped, not glued — the skin peels off and re-fits at the next site.

Cold open (0:00–0:30)

Visual: Hero drone shot. Slow pull-up-and-back, the fully skinned dome sitting in the [paddock] in low golden light. Hold on it. This is the money shot of the whole build so far — the moment it reads as a real structure, not a frame.

VO (quiet, a bit stunned):

"Eleven weeks ago this was a pile of conduit on the ground. Today... it's got a skin. It actually looks like a dome."

Visual: Hard cut to a fast montage — a panel being offered up to the frame, a screw driving, a fixing tugged to test it, a seam getting sealed. Three or four cuts, tight, rhythmic.

VO:

"But getting here meant cutting and numbering [hundreds] of little triangles. So let's talk about how you actually wrap a sphere."

On-screen: WEEK 11 — COVER title card over the last frame.

Intro to-camera (0:30–~1:30)

Visual: Stephen to-camera, standing at the base of the half-skinned dome, panels stacked and numbered behind him. Handheld, natural light.

To-camera (conversational):

"Alright — week eleven. This is the one I've been weirdly nervous about, because this is where the whole thing either looks right... or looks like a dodgy tent.
A 6V dome — that means each big edge of the icosahedron is split into six. Which sounds neat until you realise it leaves you with a LOT of triangles to cover. [Hundreds] of panels, and they're not all the same. A 6V has [several] different triangle sizes, so you can't just stamp them out and slap them on. Every single one has to go in the right spot.
So the plan: pre-cut everything, number everything, and then build the skin in rings — start at the ground, work my way up to the top. Seal the seams as I go. And leave a big gap at the front, because next week a forklift-wide doorway goes in and I do not want to be cutting a hole in a finished dome."

On-screen: simple lower-third — 6V dome = [6] subdivisions per edge → [lots of] triangles, [several] sizes

Build segments

Segment 1 — Pre-cut & number every panel (1:30–3:45)

To-camera / VO:

"Rule one of skinning a dome: do not freestyle it. If you grab a random panel and hope it fits, you'll get to ring four and discover nothing lines up.
So everything gets cut on the ground first. I've got [a few] panel types — call them A, B, C [adjust to actual] — based on the triangle sizes. Each one gets a label and a number, and they get stacked in the order they go on. It's basically a giant flat-pack of my own making."

Shot notes:

Overhead / top-down of panels laid out flat, sorted into labelled piles. Clean, satisfying.
Close-up: marker writing a number on the back of a panel.
Insert: a hand-drawn or printed numbering map / "key" taped to a board. [Confirm a physical key exists.]
Time-lapse of the cutting station — repetitive cuts, fast.

VO (honest aside):

"[I miscounted the first batch / cut a few of the wrong size] — so factor in waste. Cut spares of the small ones especially, they're the easy ones to wreck."

Segment 2 — Skinning in rings, base up (POV) (3:45–6:30)

To-camera (at the dome):

"Now the fun bit. You work in rings — finish the whole bottom row before you go up. Two reasons: the lower panels carry the ones above, and it keeps everything square as you climb. If the base ring's out by a couple of degrees, that error multiplies all the way to the top."

Visual: POV (head-mounted / chest-mounted cam) of a panel going on — offering it up, lining the edges, driving the first fixing, then the rest. Let one full panel play close to real-time so people feel the process.

VO over POV:

"Line it up with its neighbours, get one fixing in to hold it, check it sits flush, then lock it down. No lifting corners. Every edge wants to be tight to the frame — a lifting edge is where the wind gets in, and where water gets in."

Shot notes:

POV install of one panel, near real-time.
Time-lapse of a full ring going on, low to high.
Ground-level wide showing the skin "climbing" the frame across the day.
Detail: tug-testing a fixed panel — it doesn't move.
[Night shoot option: a ring going on under the work lights — ties to the livestream, looks great on camera.]

Segment 3 — Seal & weatherproof the seams (6:30–8:00)

To-camera:

"Panels on is only half the job. The seams are where a dome lives or dies. Every joint between triangles is a potential leak, so every one gets sealed."

Visual: Close-up of seam treatment — [tape / sealant / batten strip — confirm method] being run along a joint. Hands working, slow and deliberate.

VO:

"I'm going [top-down logic on water — seal so water sheds over the joint, not into it / overlap like roof tiles]. The whole point of this dome is to HARVEST water on the inside — so the last thing I want is uncontrolled water coming through the skin. Seams sealed, edges down, nothing lifting."

Shot notes:

Macro of sealant/tape going on.
Water test if time allows — [hose a sealed section, show it sheds]. Great visual proof.
Walk-around showing every seam done.

Segment 4 — Frame the entrance, leave the gap (8:00–8:45)

To-camera (standing in the opening):

"And here's the bit that takes discipline — I'm leaving a hole. This whole front section stays open. Next week a doorway goes in wide enough to drive a forklift through, so I'm framing the opening out now and leaving the gap to suit. Skin right up to the edges of it, seal those edges... and stop."

Shot notes:

Stephen standing inside the framed-out opening, gesturing the width — [reference the forklift width].
Detail: the panels stopping cleanly at the framed edge.
Tape measure / mark showing the planned doorway width. [Confirm dimension.]

The science / why it matters

VO over graphics + footage:

"Two jobs for this skin. One: shed weather. Adelaide can throw heat, wind and the odd hard downpour at this thing, and the cover has to take all of it without lifting or leaking. Two — and this is the weird one — it has to manage AIRFLOW.
This is an atmospheric water harvester. The dome works with the air inside it: [the cover helps control how air moves and where it cools so moisture can condense and be collected]. Seal it too tight and it can't breathe; leave it leaky and the weather wins. So the skin is a balance — weatherproof, but working WITH the air, not just sealing it off."

On-screen graphic: simple diagram — arrows for water shedding OFF the outside, separate arrows for managed airflow INSIDE. [Verify the actual airflow/condensation logic before final VO — this is the science the channel is built on, get it right.]

Honest beat

To-camera (tired, end of day):

"Real talk — this took way longer than I planned. [A whole ring had to come back off because I'd worked the base out of square / a batch of panels didn't fit and I had to re-cut]. There's a moment around panel [two hundred] where you stop seeing triangles and just start seeing... your own bad decisions from three weeks ago.
But." (looks up at it) "Look at it. It's a dome. With a skin. I can't quite believe it."

Visual: Hold on him looking up at the dome, then a slow tilt up the finished skin.

Outro + next-week tease

To-camera:

"So that's the cover — pre-cut, numbered, ringed up base to top, seams sealed, edges all down, and a forklift-sized hole sitting in the front waiting.
Next week we deal with that hole. The entrance goes in — a doorway you can drive machinery straight through. If the skin was the milestone where it looked finished, the door's the one where it becomes usable."

Visual: Final hero drone shot again — but now hold longer, maybe orbit. Push-in on the dark gap of the entrance.

VO (tease):

"Next week: the way in."

On-screen: NEXT WEEK — WEEK 12: THE ENTRANCE · Livestream nightly on [second channel name]

Shot list

Hero drone — fully skinned dome (cold open + outro). Golden hour, pull-back and orbit. THE milestone shot.
Fast install montage (cold open) — panel up, screw, tug-test, seal.
Intro to-camera at base of half-skinned dome.
Overhead of panels sorted into labelled piles.
Close-up: numbering a panel; the numbering key/map.
Time-lapse: cutting station.
POV: one full panel install, near real-time.
Time-lapse: a full ring going on, base to high.
Ground wide: skin climbing the frame over the day.
Detail: tug-test, no lifting edges.
[Night shoot: ring install under work lights.]
Macro: seam sealing method.
[Water test on a sealed section.]
Stephen inside framed entrance opening, showing width.
Detail: panels stopping clean at the doorway frame.
Honest-beat to-camera, end of day, tilt up the skin.
Outro drone push-in on entrance gap.

On-screen text / graphics

WEEK 11 — COVER (title card)
Lower-third: 6V dome = [6] subdivisions per edge → [lots of] triangles, [several] sizes
Panel-key graphic / numbering map callout
"Work in RINGS — base up" simple animated callout
Science diagram: water sheds OFF outside / airflow managed INSIDE
NEXT WEEK — WEEK 12: THE ENTRANCE
Persistent: Atmosfarm · @atmosfarm · Adelaide, South Australia lower-corner bug
Nightly livestream → [second channel]

Runtime breakdown

Section	Start	Duration
Cold open	0:00	0:30
Intro to-camera	0:30	1:00
Seg 1 — Pre-cut & number	1:30	2:15
Seg 2 — Skin in rings (POV)	3:45	2:45
Seg 3 — Seal seams	6:30	1:30
Seg 4 — Frame entrance	8:00	0:45
Science / why it matters	8:45	0:45
Honest beat	9:30	0:20
Outro + tease	9:50	0:25
Total		~10:15

[Trim Seg 1 or Seg 2 by ~15s each to land exactly on 10:00 if needed.]

↑ top

Atmosfarm — Weekly Build Video 12: Entrance

Build week: Week 12 of [14] · Runtime: ~10 min · Status: PLACEHOLDER (draft VO/to-camera + shot direction — refine later)

Logline

The dome gets its mouth: we frame and hang a forklift-wide door, weather-seal the threshold, and for the first time the whole thing is closed in. One door turns a frame into a building.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): The door's a bolted frame in the opening — it comes off with the rest.

Cold open (0:00–0:30)

Shot notes: Slow push-in on the open doorway at dawn — the dome still gaping, wind moving through it. Cut to a hard, satisfying SLAM of the finished door at the end (flash-forward), then back. Sound design carries this: wind howling through the open shell, then a clean thunk and latch click, then near-silence inside.

VO (low, over the open doorway):

For twelve weeks this thing has been open to the sky. Wind straight through it. Rain straight through it. Birds, leaves, the lot.
Today that stops. Today it gets a door.

[Cut to the slammed/latched door — 1 second — then title card.]

On-screen title: ATMOSFARM — WEEK 12: ENTRANCE

Intro to-camera (0:30–~1:30)

Shot notes: Stephen standing in the open doorway, one hand on the conduit frame, dome behind him. Handheld, conversational. Mid-morning light. [Adelaide / build site.]

To-camera:

Right — so this is the bit where it finally feels like a building instead of a... very ambitious tent frame.
The plan today: build the door, fit it to the opening, hang it so it actually swings true and latches — not just leans there looking pretty — then seal the threshold and the whole perimeter so weather doesn't sneak in around the edges.
Now — the thing you'll notice straight away is this opening is big. Like, deliberately, awkwardly big. That's not a mistake and it's not me being lazy with the measurements. This doorway is wide enough to drive a forklift through. On purpose. I'll get into why that matters, because it changes basically everything about how this dome works as an actual working space — not just somewhere you walk in and out of.
And the other big thing: once this door's on, the dome is sealed. Closed in. Which means the airflow envelope — the thing the whole water-harvesting idea depends on — that's got to actually hold. So we're not just hanging a door, we're closing a loop.
Let's build it.

On-screen text: Today — frame · hang · seal · close it in

Build segments

Segment 1 — Building & fitting the door to a wide opening (1:30–4:00)

Shot notes: Workbench / cut-and-weld montage. Tape measure across the opening (hold the wide number on screen). Squaring the frame on the ground. Test-offering the door up into the opening — the inevitable "doesn't quite fit first go" moment. Timelapse the fabrication, drop back to real-time for the fit checks.

VO / to-camera (mix):

A normal door you buy off a shelf. This one I'm building, because there's no off-the-shelf anything for an opening this size in a dome wall.
[Door build assumption: timber/steel hybrid frame, [cladding type TBC] — confirm materials in edit.] The frame has to be dead square or it'll bind when it swings, so I'm checking diagonals corner to corner — if those two measurements match, it's square.
Offering it up... and — yeah. Course it doesn't fit first go. It never does. [Trim/adjust on camera — show the problem, show the fix.]

To-camera (honest aside, quick):

The hard part with a dome is nothing's a flat rectangle. The wall curves away from you. So a flat door into a curved-ish frame means you fight the geometry the whole way. [Note in edit: explain how the opening was framed flat/plumb to give the door a square home.]

On-screen text: Forklift-wide = [~X.X m] across — confirm measurement

Segment 2 — Hanging it: swing true & latch (4:00–6:00)

Shot notes: Hinges going on — close-up on the heavy-duty hardware ([gate-grade hinges for the weight/width — confirm]). The first swing. The moment it actually closes and the latch catches. Get the sound of the latch clean. A wide shot of the door fully open showing the forklift-clear span.

VO / to-camera:

A door this wide is heavy, and weight plus width is the enemy of a door that swings true. Too much sag and the bottom corner drags, or it won't latch. So the hinges are oversized — [gate-grade, TBC] — and I'm hanging it [in stages / with a helper / on packers] so I can dial the gap before I commit.
First swing... [pause for it on camera].
And — latch. There it is. Listen to that. (latch click) That's the noise of a building that closes.

To-camera (standing in the open span):

And look at this — door fully open, that is a clear run straight into the middle of the dome. You could drive a pallet of soil, a tank, a full garden bed, whatever — straight in. No carrying it through a person-sized door one armful at a time. That's the forklift thing again, and next week you'll see exactly why I wanted it.

Segment 3 — Sealing the threshold & perimeter (6:00–7:45)

Shot notes: Close on the threshold — weather strip / seal going down. Perimeter brush seals or compression seal around the frame. The "smoke / paper test" gag if it films well — hold a bit of paper or a smoke source near the closed edge to show airflow (or lack of it). Then a step/access path being finished at the entrance — a poured/laid step or ramp.

VO / to-camera:

A door that closes isn't the same as a door that seals. There's a difference, and the whole project lives in that difference.
Threshold seal along the bottom — [type TBC]. Perimeter seal around the frame so the edges aren't leaking air. [Show seal product / method.]
And because forklift-wide means a big threshold to cross, I'm finishing the access into it — [a step / a low ramp so you can wheel or drive over it cleanly — confirm].

To-camera (the test):

Quick test — [paper / smoke] at the edge with the door shut. If it's moving, I've got a leak to chase. If it sits still... we're sealed. [Capture result — keep it honest either way.]

On-screen text: Sealed ≠ closed. We need sealed.

The science / why it matters

Shot notes: Simple animated overlay on a clean shot of the closed dome — arrows showing the controlled airflow envelope: intended air path in/through/out, vs. uncontrolled leaks at a bad seal. Keep it rough-and-clear, not glossy.

VO:

Here's why a door is more than a door on this build.
The dome harvests water out of the air. To do that you want control over the air inside it — how it moves, its moisture, its temperature. The second you've got random leaks around a badly hung door, you've lost that control. The air you've carefully conditioned just bleeds out the gaps, and the moisture goes with it. [Confirm exact mechanism against the water-harvesting design notes.]
So sealing the perimeter isn't trim-work. It's the difference between a dome that manages its internal air and one that just... leaks. Closing this door is the moment the airflow envelope becomes a real, closed system.
And the forklift-wide opening — that's the working-dome trade-off. A massive door is harder to seal and harder to hang. But a sealed small door means you'll never get a tank, a bed, or a future bit of equipment inside without dismantling something. So you build it big, then you do the hard work of sealing big. Capability first, then make it tight.

On-screen text / graphic: Controlled airflow envelope — sealed door = closed system

Honest beat

Shot notes: End of day, lower light. Stephen sitting in the doorway threshold, door beside him. Quiet, real. [Use whatever genuinely went wrong today.]

To-camera:

Not going to pretend that was smooth. [Placeholder for the real struggle: e.g. the door bound on the first hang / the seal gap was worse than I thought / I had to re-cut the frame.] A door this wide punishes every millimetre you're out.
But — (taps the closed door) — it's on. It shuts. And standing inside just now with it closed for the first time, it went quiet. Properly quiet. After twelve weeks of working in something that's basically outside, that hit me a bit. It's a room now.

Outro + next-week tease (≈8:30–10:00)

Shot notes: The big to-camera walk-through of the now-sealed dome — Stephen opens the door from outside, steps in, closes it, walks the interior. Interior establishing shots now it's closed in — light through the frame, the enclosed volume. Then OUTSIDE and UP: drone progress shot rising off the dome, now visibly a complete sealed structure on the site. Tease shots: empty interior floor where the beds will go.

To-camera (walk-through):

Come in. (opens, steps through, closes door behind) That's it — sealed. Whole dome, closed in. (walks the interior, gestures around) Twelve weeks ago this was a slab and an idea. Now it's an enclosed space you can shut the door on.
(stops, looks at the bare floor) And it's empty. Which is the whole point of next week — because that forklift-wide door I keep banging on about? Next week we use it. The garden beds go in. We start turning this from a structure into something that actually grows.
Livestream's on every night on [livestream channel] if you want to watch it happen between these videos. I'll see you in the next one.

On-screen text: NEXT WEEK — Garden beds go in On-screen text: Live every night → [livestream channel handle — confirm]

Shot list

Cold open: dawn push-in on open doorway, wind through shell; flash-forward door SLAM + latch.
Sound: wind-through-open-shell, latch click, interior near-silence once sealed.
Intro to-camera in the open doorway, hand on frame.
Fabrication montage/timelapse: cutting, squaring, welding/assembling the door.
Tape measure across opening — hold the wide measurement.
Diagonal square check, corner to corner.
Test-offer door into opening (the "doesn't fit first go" beat) + trim/fix.
Hinge close-ups (heavy-duty hardware).
First swing; first successful latch (clean audio).
Wide: door fully open showing forklift-clear span into dome centre.
Threshold seal laid; perimeter seal around frame.
Air-leak test (paper/smoke) at closed edge.
Step/access path finished at entrance.
Animated airflow-envelope overlay on closed dome.
Honest beat: Stephen in threshold, low light.
Walk-through: open from outside → step in → close → walk interior.
Interior establishing shots, now closed in.
Drone progress shot rising off completed sealed dome.
Tease: bare interior floor where beds will go.

On-screen text / graphics

Title: ATMOSFARM — WEEK 12: ENTRANCE
"Today — frame · hang · seal · close it in"
"Forklift-wide = [~X.X m] across" [confirm measurement]
"Sealed ≠ closed. We need sealed."
Airflow graphic: "Controlled airflow envelope — sealed door = closed system"
"NEXT WEEK — Garden beds go in"
"Live every night → [livestream channel handle]" [confirm]
Lower-third: Stephen James Lovegrove · @atmosfarm · Adelaide, South Australia [confirm style]

Runtime breakdown

Section	Start	Length
Cold open	0:00	0:30
Intro to-camera	0:30	1:00
Seg 1 — Build & fit door	1:30	2:30
Seg 2 — Hang: swing & latch	4:00	2:00
Seg 3 — Seal threshold & perimeter	6:00	1:45
The science / why it matters	7:45	0:45
Honest beat	8:30	0:35
Outro + next-week tease	9:05	0:55
Total		~10:00

PLACEHOLDER draft — all [bracketed] items (materials, hardware specs, exact opening width, livestream handle, the real day's struggle for the honest beat) to be confirmed and the VO tightened against actual footage in the edit.

↑ top

Atmosfarm — Weekly Build Video 13: Garden Beds

Build week: 13 of 14 | Runtime: ~10 min | Status: PLACEHOLDER (draft VO/to-camera + shot direction — refine later)

Logline

The week the dome stops being a structure and starts being a system. Beds go in, get tied into the aquaponics loop from Week 10, the irrigation gets tested for real, media goes down, and the first seedlings hit the soil. By the end of the week the dome is, for the first time, alive.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): Beds are on skids and roll out the forklift-wide door — lift and load.

Cold open (0:00–0:30)

Shot notes: No talking yet. Open on a single seedling tray sitting on bare ground inside the dome — early light through the panels, dust in the air. Slow push-in. Cut to hands lifting the first seedling out of the tray. Cut to the empty bed waiting. Hold a beat on the gap between the two.

VO (low, over the visuals):

"Thirteen weeks. A doorway, a frame, a tank, a loop of water with nothing growing in it yet. [pause] Today that changes."

On-screen text (lower third, small): WEEK 13 — GARDEN BEDS

SFX: ambient dome tone, birds outside, the faint hum of the pump from the aquaponics loop already running.

Intro to-camera (0:30–~1:30)

Shot notes: Stephen to-camera, standing inside the dome between the empty beds. Handheld, slightly loose. The planted interior is NOT visible yet — that's the payoff later. Keep it conversational, not scripted-feeling.

To-camera (draft):

"Alright — week thirteen. We're nearly there. [gestures around] Last week we put the forklift-wide doorway in, which means everything I'm doing today, I can actually get straight in here — beds, growing media, all of it, no carrying buckets through a gap the size of a letterbox.
The plan for the week: build the beds, get them positioned to the layout I've had in my head for [months / since day one], plumb them into the aquaponics loop we put in back in Week 10 — so the water's already circulating, it just hasn't had anywhere to go yet. Then test the irrigation and the drainage properly, fill and level the media, and — the bit I've actually been waiting for — get the first plants in.
If it all works, by the end of this video the dome is alive. That's the goal. Let's go."

On-screen text: THE PLAN — BEDS · LOOP · TEST · MEDIA · PLANTS

Build segment 1 — Building & positioning the beds (1:30–3:45)

To-camera / VO (draft, mix of on-site to-camera and VO over build B-roll):

"Beds first. [Material assumption: timber / Colorbond / wicking-style beds] — I went with [X] because [Y reason: durability / cost / suits the aquaponics flow].
The layout matters more than people think. This is a dome — it's a circle, not a rectangle — so the beds have to follow the curve and still leave me paths I can actually walk and work. [Describe layout: e.g. radial beds around a central feature / horseshoe / tiered]."

Shot notes:

Time-lapse of beds being assembled outside or in the staging area.
Forklift / trolley bringing a bed straight through the wide doorway — call back to last week. Get the satisfying "it fits" moment.
Top-down or high-angle (ladder/interior drone) showing beds dropping into position around the dome floor — the layout revealing itself.
To-camera moment kneeling by a bed, tapping it level, checking spacing.

On-screen text: [BED TYPE] · [N] BEDS · LAYOUT FOLLOWS THE CURVE

Build segment 2 — Tying into the aquaponics loop (3:45–6:00)

To-camera / VO (draft):

"Now the part that makes this a system instead of just garden beds. Back in Week 10 we plumbed the aquaponics loop — water's been circulating from [the tank / fish side] all this time with nowhere productive to go. Today the beds become the destination.
[Explain the connection in plain terms: water in from the loop here, through the media / grow bed, drains back out here and returns to the tank.] The whole point is it's a closed loop — the plants clean the water, the water feeds the plants, and I'm not pouring fresh supply in every day. That's the atmospheric-water philosophy of the whole build, just at bed scale."

Shot notes:

Close-ups: connecting feed lines into the beds, fittings going on, the junction where the loop meets the bed.
The "first flow" moment — water actually running INTO a bed for the first time. This is a hero shot. Slow-mo if it lands.
Cutaway to the pump / tank end of the loop so the viewer sees both ends of the circuit.
[If night work happens this segment: use the night lighting rig — water moving under the work lights looks great and reinforces the round-the-clock build.]

On-screen text (simple graphic): TANK → BEDS → DRAIN → TANK (closed loop)

Build segment 3 — Test, fill, level, and the first plants (6:00–8:30)

To-camera / VO (draft):

"Before any media goes down — test. Run the irrigation, watch the drainage, find the leaks now and not after there's [X kg] of media sitting on top of it. [Note any actual problem found here — a fitting that wept, a bed that drained slow, whatever really happens. Keep it honest.]
Once I'm happy it holds and it drains — media in. [Media type assumption: clay pebbles / scoria / coco / soil blend] — straight in through the wide door, fill, level it off. [Describe leveling so plants sit right relative to the waterline.]
And then — finally — plants. [What's going in first: leafy greens / herbs / seedlings raised earlier in the series?] First seedlings into the first working bed in the dome."

Shot notes:

Irrigation test: water moving across/through a bed, drainage running clear. Show a genuine check — hand in the flow, watching the drain.
Media fill time-lapse — bringing it in through the doorway, dumping, raking level.
THE moment: hands placing the first seedling, firming it in. Get this clean and close. Maybe a wide right after — small green plant, big empty dome around it.
Reverse / reaction: Stephen standing back looking at the first planted bed.

On-screen text: TEST FIRST · MEDIA: [TYPE] · FIRST IN: [CROP]

The science / why it matters (8:30–9:00)

VO over clean B-roll of the planted, working interior:

"Here's why this week is the turning point. Up to now the dome's been catching and storing water — but stored water is just inventory. The beds are where it becomes food. The aquaponics loop means the same water cycles again and again: nutrients from [the fish / nutrient side] feed the plants, the media and roots filter it, and it returns cleaner than it left. In a place like [South Australia], where water is the whole question, growing food on a closed loop fed largely by water you pulled out of the air — that's the entire thesis of Atmosfarm, sitting in a bed you can touch."

Shot notes: Slow, considered B-roll. The dome with green in it. Light through panels onto leaves. Water moving in the loop. This is the most "cinematic" the episode gets — let it breathe.

On-screen text (optional graphic): WATER STORED → WATER WORKING → FOOD

Honest beat (9:00–9:20)

To-camera (draft — keep it real, not staged):

"I'll be straight — [honest issue placeholder: the layout I'd drawn didn't survive contact with an actual circle / one bed drained way slower than I wanted / I planted later in the day than I meant to and I'm not 100% these seedlings love me yet]. That's the build. It's not finished and it's not perfect. But there's something growing in here now that wasn't this morning, and after thirteen weeks of dirt and steel — that hits different."

Shot notes: Tired-but-pleased energy. Maybe sitting on the edge of a finished bed. Golden hour or the night rig.

Outro + next-week tease (9:20–10:00)

To-camera + VO (draft):

"That's week thirteen. The dome's alive. [beat]
Next week is the last one. Week fourteen — the finale. We get the exhaust sorted, do the landscaping around the outside, and wrap up the hundred days. Everything we've built, finished and standing. I genuinely can't believe we're here.
If you've been following along — the nightly livestream's still running over on [livestream channel], come hang out while I'm actually in there working. And come back next week for the end of it. Cheers."

Shot notes:

To-camera goodbye inside the planted dome.
Final shot: interior drone / high progress shot pulling up and back through the dome — first time we see the WHOLE planted interior in one frame. Tease title card for the finale.
Hard cut to black / logo.

On-screen text: NEXT WEEK — FINALE: EXHAUST · LANDSCAPING · 100 DAYS | SUBSCRIBE · LIVESTREAM NIGHTLY @ [CHANNEL]

Shot list

#	Shot	Type	Notes
1	Lone seedling tray, light through panels	Cinematic / push-in	Cold open
2	First seedling lifted from tray	Close-up	Cold open
3	Empty bed waiting	Hold	Cold open contrast
4	Stephen intro between empty beds	To-camera, handheld	Don't reveal planted interior
5	Bed assembly	Time-lapse	Staging area
6	Bed through wide doorway	Action	Week 12 callback — "it fits"
7	Beds dropping into layout	High angle / interior drone	Layout reveal
8	Leveling / spacing check	Close-up to-camera
9	Loop connection / fittings	Macro close-up
10	First water INTO a bed	Hero / slow-mo
11	Pump + tank end of loop	Cutaway	Show both ends
12	Irrigation + drainage test	Action, hand in flow	Honest check
13	Media fill through doorway	Time-lapse
14	First seedling planted	Close-up	THE moment
15	Wide: small plant, big dome	Wide
16	Planted interior B-roll	Cinematic	Science segment
17	Honest-beat to-camera	To-camera, on bed edge
18	Night-rig water/work shot	Optional night	Round-the-clock build
19	Outro to-camera	To-camera
20	Full planted interior pull-up	Interior drone	Finale tease

On-screen text / graphics

WEEK 13 — GARDEN BEDS (cold open lower third)
THE PLAN — BEDS · LOOP · TEST · MEDIA · PLANTS
[BED TYPE] · [N] BEDS · LAYOUT FOLLOWS THE CURVE
TANK → BEDS → DRAIN → TANK (closed-loop graphic)
TEST FIRST · MEDIA: [TYPE] · FIRST IN: [CROP]
WATER STORED → WATER WORKING → FOOD
NEXT WEEK — FINALE: EXHAUST · LANDSCAPING · 100 DAYS
SUBSCRIBE · LIVESTREAM NIGHTLY @ [CHANNEL]

Runtime breakdown

Segment	Start	End	Duration
Cold open	0:00	0:30	0:30
Intro to-camera	0:30	1:30	1:00
Seg 1 — Building & positioning beds	1:30	3:45	2:15
Seg 2 — Tying into aquaponics loop	3:45	6:00	2:15
Seg 3 — Test, fill, level, first plants	6:00	8:30	2:30
The science / why it matters	8:30	9:00	0:30
Honest beat	9:00	9:20	0:20
Outro + next-week tease	9:20	10:00	0:40
Total			10:00

↑ top

Atmosfarm — Weekly Build Video 14: Exhaust & Landscaping

Build Week 14 · ~10 min · Status: PLACEHOLDER · SERIES FINALE

Logline

The last build week. We fit the exhaust, close the air loop, and watch the whole machine breathe for the first time — then a crew of volunteers turns raw ground into a finished site. One hundred days, and the dome is done.

Design rule — temporary by design

Recurring beat (work into a to-camera line this episode): The exhaust unbolts and the tank drains — the only thing that stays is the bare ground.

What's next (no ask): it works, and it all comes apart and moves. This was always temporary by design and self-funded — proof first. Down the track Atmosfarm finds permanent ground and goes in to stay; that's the vision, not a pitch. (This is the tone for the launch 'Vision / What's Next' video — show the dream, don't sell.)

Cold open (0:00–0:30)

VISUAL: Hard cut, no music. Tight on the exhaust outlet. A strip of survey tape or a feather held near it. Dead still. Then — the system kicks on and the tape snaps sideways and holds, fluttering hard. Hold on it.

VO (low, to-camera over the top):

"Intake's been pulling air for weeks. The guts have been making water for weeks. But the loop's never actually been closed. Today it does — and if it works, this thing breathes on its own for the first time."

SMASH CUT to title card over the hero drone shot (see below).

ON-SCREEN: WEEK 14 — THE LAST ONE

Intro to-camera (0:30–~1:30)

VISUAL: Stephen on site, dome behind, ground still rough/torn up from the build. Honest framing — tools everywhere, it's not pretty yet. Late afternoon light.

To-camera (conversational):

"Right. So this is it — week fourteen, the last build week before we go live. And there's basically two jobs left, and they're both big.
One: the exhaust. Up to now the air system's been half a system — air goes in, water gets pulled out of it, but the spent air's had nowhere proper to go. Today we build and fit the venting so the whole loop is complete — intake, water-2-air, exhaust — and then we run the full draw end to end. First time the entire thing works as one machine.
Two: the site itself. [Number] volunteers are turning up this week to help with the landscaping — grading, paths, planting, the lot. Working-bee energy. I genuinely could not get this done in time on my own, so a lot of what you'll see this week is other people's hands.
And because we've got the night-work lights rigged, the last push can run well into the evening. So let's go."

Build segment 1 — Building & fitting the exhaust (1:30–3:45)

Draft to-camera / VO:

"The exhaust isn't glamorous but it matters more than it looks. The air coming out the back of the water-2-air stage is colder and it's had a lot of its moisture stripped — we need to get it out and away cleanly so it's not just getting sucked straight back into the intake and short-circuiting the whole loop. So placement is half the job here."

VO over build montage:

"Cutting the ducting to length. Fitting the outlet. Sealing the joints so we're not losing draw to leaks." [confirm exhaust material — ducting / conduit-matched / fabricated cowl]

Beat — a small problem (keep it honest):

"First fit, [the outlet sat too close to the intake side / a join leaked / clearance was tight] — so [reposition / re-seal / re-cut]. Builder problems. You don't get it right first go." [adjust to actual issue]

SHOT NOTES: Close-ups of cutting and sealing. One clean wide of the exhaust in place against the dome. Over-the-shoulder of the join being checked. Capture the leak/fix beat if it happens — don't stage it.

Build segment 2 — Closing the loop & first draw (3:45–5:15)

"With the exhaust on, every part of the air path is now physically connected. Intake at one end, water-2-air in the middle, exhaust at the other. So before we landscape over anything, we run it."

(This rolls into the prototype test — see below.)

SHOT NOTES: Walk the full air path on camera, hand following the route, naming each stage as you pass it. This doubles as a recap of the whole build for new viewers landing on the finale.

Build segment 3 — Landscaping with the volunteers (5:15–7:15)

VISUAL: This is the community heart of the episode. Volunteers arriving, gloves on, wheelbarrows, the site transforming. Energy, chatter, a bit of laughing. Let it breathe — fewer words, more faces.

VO (warm, light):

"And then the cavalry showed up."

"Final grading first — getting the ground sitting right around the base. Then the paths in. Then planting out the exterior. [Number] people who didn't have to be here, here anyway, on a [weekend / weekday] giving up their time to get a stranger's dome over the line."

Optional volunteer soundbites: [grab 2–3 short to-camera lines from volunteers — why they came, what they think of it.]

"We worked till the light went. Then we kept working, because — night lights." [cut to night landscaping under the work lighting]

SHOT NOTES: Drone up over the working-bee mid-afternoon — bodies moving around the dome. Ground-level candids. The before/after of one path or bed. Night work under the rigged lights, headlamps, breath in the cold if it's cold. Get names for credits. [confirm volunteer count + day]

Build segment 4 — Snag list & sign-off (7:15–8:00)

"Last job of the whole build: the snag list. Walking the entire thing with a notepad, finding every little unfinished bolt, every gap, every 'I'll come back to that' — and actually coming back to it."

VISUAL: Stephen with a clipboard / phone, ticking items, tightening last fixings. Final wide of the finished site, paths in, plants in, ground clean.

"And... that's it. That's the build. Signed off."

SHOT NOTES: Pencil ticking a list. A genuinely final fixing being tightened. Slow push-in on the completed dome.

The science / why it matters — completing the air loop (8:00–8:40)

VO over simple animated diagram:

"Here's the whole thing in one picture. Ambient air gets drawn in at the intake. It passes through the water-2-air stage, where [the moisture in that air is condensed and collected — confirm mechanism: cooling/condensation surface]. The now-drier, cooler air gets pushed out the exhaust — placed so it can't loop back into the intake.
Air in, water out, air out. That's the engine. Everything else — the dome, the structure, the site — exists to hold that loop and run it [continuously / on cycle]. Close the loop and you've got a machine that pulls drinking water out of the sky." [verify all science claims before publish]

ON-SCREEN: Animated loop diagram: INTAKE → WATER-2-AIR → EXHAUST, arrows flowing, water droplet split off at the middle stage.

Prototype test — the full end-to-end draw (8:40–9:15)

THIS IS THE PAYOFF OF THE WHOLE SERIES.

VISUAL: Everyone clears back. Hand on the switch. Power on.

To-camera (real, unscripted as possible):

"Okay. Full system. Everything connected for the first time. Here goes."

VO over the test:

"[Confirm what we're measuring — airflow at exhaust / draw at intake / first condensate]. Watching the [tape / anemometer / flow indicator] at the exhaust... and there it is. The whole loop, drawing, end to end." [capture the real reading + reaction]

"[If first water appears on camera — lead with it. If not, note that draw is confirmed and water collection is the slow part / runs on the livestream.]" [adjust to actual result]

SHOT NOTES: Multi-angle if possible — intake, mid-stage, exhaust — to show the loop live at once. The genuine reaction is the shot. Do not over-cut it.

Honest beat — finishing the 100 days (9:15–9:40)

VISUAL: Night. Work lights off, dome quiet, maybe lit softly from inside. Stephen sitting, knackered. No music or very little.

To-camera (quiet, real):

"One hundred days. [Adjust to actual figure.]
I'm not going to pretend it went to plan — heaps of it didn't. [Reference a real low point — the week that nearly broke it.] There were days I genuinely didn't think this would be standing.
But it is. And it works. And honestly the part I didn't expect is that the last week — the bit that got it over the line — wasn't me. It was everyone who turned up. So. Thank you. That's the build done."

[Keep this short and un-polished. Let it be a bit awkward. Don't sand it down.]

Outro + next-week tease — the bridge to launch (9:40–10:00)

VISUAL: Hero drone fly-over — the finished, landscaped site from above, golden hour. The shot that's also going in the 100 Days film.

To-camera / VO:

"The build is finished. But Atmosfarm isn't a build — it's a thing that's meant to run. So next: we go live.
The nightly livestream keeps going on the other channel — come watch it actually make water in real time. And the full story of all hundred days — the '100 Days' film — drops on launch. [Confirm launch date — Oct 14 2026.]
Thanks for building it with me. See you on launch day."

ON-SCREEN: THE BUILD IS DONE. → LAUNCH: [OCT 14 2026] → 100 DAYS — coming + livestream channel handle + subscribe prompt.

Shot list

Cold open: static exhaust outlet w/ tape → snaps to life (the breath shot)
Hero drone fly-over of finished landscaped site, golden hour (DUAL USE: this episode + timelapse + 100 Days film) — PRIORITY GRAB
Exhaust build: cutting, fitting, sealing joints (close-ups)
Exhaust in place against dome (clean wide)
Leak/fix or repositioning beat (if it happens — real, unstaged)
Walk-the-full-air-path tracking shot (intake → mid → exhaust)
Volunteers arriving / working-bee wide (drone + ground)
Volunteer candids + 2–3 soundbites to camera
Before/after of a path or planting bed
Night landscaping under rigged work lights (headlamps, cold breath if cold)
Snag list: clipboard ticking, final fixing tightened
Prototype test: multi-angle live loop + genuine reaction
Honest beat: night, dome softly lit, Stephen seated and tired
Science diagram (motion graphic, post)
Closing dome hero (push-in)

On-screen text / graphics

WEEK 14 — THE LAST ONE (title)
Lower-third stage labels during air-path walk: INTAKE / WATER-2-AIR / EXHAUST
Animated air-loop diagram (INTAKE → WATER-2-AIR → EXHAUST, water droplet split-off)
Test readout overlay: [airflow / draw figure — confirm metric]
Day [X] of 100 counter
Volunteer name credits (gather on the day)
THE BUILD IS DONE.
LAUNCH: [OCT 14 2026]
100 DAYS — coming + livestream handle + subscribe

Runtime breakdown

Section	Start	Duration
Cold open	0:00	0:30
Intro to-camera	0:30	1:00
Build 1 — Exhaust build & fit	1:30	2:15
Build 2 — Close the loop & first draw	3:45	1:30
Build 3 — Landscaping w/ volunteers	5:15	2:00
Build 4 — Snag list & sign-off	7:15	0:45
Science — completing the air loop	8:00	0:40
Prototype test — full end-to-end draw	8:40	0:35
Honest beat — finishing 100 days	9:15	0:25
Outro + launch tease	9:40	0:20
Total		~10:00

PLACEHOLDER — refine after the build week. Verify all science claims, the day count, exhaust materials/mechanism, volunteer count + day, and the live test result/metric before publish. Bracketed items = assumptions to confirm.

↑ top