BUILT FOR STORYTELLERS · EARLY ACCESS

Stand in the storybefore you tell it.

Take Zero turns any real place into a set you can stand in. Walk it, shape camera moves like sentences, and show your team exactly what you see.

Get the beta

LOCAL-FIRSTENCRYPTED ROOMSRUNS ON YOUR GPUYOURS FOREVER

Why this exists

You already see it.

The push through the doorway. The wide that finally opens the scene. The low angle nobody will sign off on rigging. You've been directing this thing for weeks, and the only place it exists is in your head.

Take Zero gives the picture in your head a place to stand. Walk the location tonight. Try the move that might not work. Find the frame that does. And when someone asks what you mean, stop describing it. Show them.

Real places, faithfully. Nothing added that wasn't there.

What you walk away with

Six things you can't easily get any other way.

Problems, found early

Blocked sightlines, tight corners, impossible rigs, discovered weeks out, not with the crew watching you improvise.

Moves you can sculpt

Every camera move is a path. Bend it, retime it, replay it until it feels inevitable. Then keep it forever.

A cut before the shoot

Run every camera live, call the cuts, and leave with an actual edit on an actual timeline, not a folder of screenshots.

The room, together

Director on a couch, DP in another city. Same set, same moment, walking it while you talk it through.

A plot that plays

Every setup lands on a real overhead plot: positions, lenses, coverage. Share it live and the moves play out on the plan.

Nothing held hostage

Your hardware, your files, no uploads. No service standing between you and your own work.

HOW IT ALL WORKS →

Multicam · live cut · timeline

Call the show before there's a show.

Every camera in the scene runs live, at once. Cut between them like a TD, by click or the number keys, while the moves play. And your cuts don't evaporate: they land on an editing timeline you can trim, razor, rearrange, and export as a film. We haven't found another tool, at any price, that takes you from blocking to a working edit in the same room.

Movement

Moves are drawn, not keyed.

Camera movement in Zero is path-based: a line you sculpt through space. Bend it, ease the ends, even the speed with one press, and it replays exactly, every take. Characters walk their own paths too, and cameras can track them. It's the difference between describing a dolly and conducting one.

A MOVE, DRAWN IN LIGHT · REPLAYED FOREVER

Rooms

Everyone on the set, at once.

Open a room and the team steps into the scan with you. Everyone walks free, everyone frames; you can watch through anyone's eyes while they drive, and shared multicam runs live for the whole room, their moves in your cells, in real time. Talk through the shots the way you would on the floor. All of it end-to-end encrypted, because your set is nobody's business but yours.

LOC_001.SOG · 84 MB · 1,412,006 SPLATS

Scout it once. Shoot it forever.

Locations fall through. Light changes. Permits expire. The scan doesn't care. Walk it at midnight from your kitchen, send it to the DP in a text, and run the shot a hundred times before the day.

The camera plot

A camera plot that plays.

Zero draws a real overhead plot straight from your setups: every position, lens, and coverage angle, with your actors marked where they stand. Print it if you want one on the wall. But the version that changes the conversation is the live one: send a link, and your team watches the moves travel the plan in real time. They scrub the timing, follow the dolly, and see the coverage, instead of squinting at a static diagram and taking your word for it.

An overhead camera plot of a basement location: the scanned floor plan in grey, five numbered orange camera positions with coverage wedges and camera moves, two marked actors, and a camera list with lenses down the right side.

BASEMENT · PLOT 001 · 5 SETUPS · 28-85MM · SHARED LIVEPLOTTED IN ZERO

Where it's heading

The scan becomes the set.

Zero today is blocking, framing, and the cut. What's coming makes the scan itself do more of the work:

Light IN DEVELOPMENT

Relighting the real room. Move the hour, shape the practicals, take your noon scan to blue hour before a single lamp is rented.

Render

The road leads to final-pixel looks: a render path that turns blocking into imagery you can put in front of a client without apologizing.

Import everything

More formats, more sources. If it can be captured, Zero should be able to stand in it.

Who it's for

Made for the people who decide the frame.

Narrative & film

The location falls through on Tuesday and you shoot Friday. Walk the new one tonight.

Commercial & branded

The client wants to see the shot, not hear about it. Send the exact frame.

Music video

One scout pass with a phone becomes every setup you'll shoot next week.

Live & broadcast

Rehearsal time is the most expensive clock there is. Spend it having already decided.

Trust

Private by architecture.

LOCAL-FIRST

Scenes are built and stored on your machine. There is no upload step to skip. It doesn't exist.

E2E-ENCRYPTED ROOMS

Collaboration is end-to-end encrypted. The room key never leaves your team; we relay ciphertext and nothing else.

YOUR GPU

Builds run on your hardware. No cloud queue, no per-render bill, no service standing between you and your work.

Your story already has a place. Go stand in it.

Get the beta

How it works

From footage to a place you can direct.

Two quiet apps, one pipeline. Intake builds the place from your footage. Zero is where you stand in it and decide.

Film it

A slow, steady walk through the place with any camera. A phone is plenty. Aim for a few minutes of coverage with lots of overlap. No rig, no markers, no technique beyond patience.

Intake builds it

Frames are pulled, poses are solved, and the scene trains on your GPU, typically well under an hour. You level the floor, crop what you don't need, and press Finish. Everything happens on your machine. The full reconstruction stack lives on the Intake page.

Zero is the set

Open the scene and walk it in first person. What you do next is the craft: cameras, lenses, moves, marks. Every decision is saved with the scene.

Inside Zero

The whole instrument.

Virtual cameras

Real focal lengths, real aspect ratios. Place as many as the scene needs and name them like you would on a truck.

Paths

Every move is a sculpted, replayable path: constant-speed timing, eased ends, one press to even the whole move out.

Multicam, live

Every camera at once, playing. Cut by click or number keys, exactly like calling a show.

The timeline

Your live cuts land on an editing timeline. Razor, trim, rearrange, scrub, undo, then export the film.

Characters on paths

A scale-true human figure that walks its own path, and cameras can track it through the scene.

Rooms

The team joins from a link into the same scan. Everyone walks, everyone frames, live, end-to-end encrypted.

Watch

See through any collaborator's camera while they drive. The fastest way to be shown, not told.

Frames, film & the plot

Stills for the deck, MP4s of the moves and the cut, and an overhead camera plot you can print, or share as a live link where the moves play.

Bring your own build

Got a set model from your production designer, or built one yourself? Import GLB or glTF and block it exactly like a scanned location.

Update in place

New capture of the same place? Swap the scan underneath and every camera, path, and cut survives.

Under the hood

A quiet surface on serious machinery.

Zero runs captured places and built sets in the same viewport, in real time, on ordinary hardware, with none of the machinery showing.

Scans and built sets

A captured place arrives as a radiance field, where glass, foliage, and soft light behave like the footage. A designed set arrives as GLB or glTF. Both open in the same scene, at the same scale.

True lens math

Virtual cameras use real focal-length and sensor geometry so 35mm in Zero frames like 35mm on the day.

Real-time by sorting

Splats are depth-sorted and streamed every frame with level-of-detail, so a laptop walks a million-splat room smoothly.

Zero-knowledge rooms

The room key derives from your password in the browser: PBKDF2 to AES-GCM. The relay moves ciphertext it cannot read, ever.

Local by architecture

Renderer, files, and builds live on your machines. The network is optional, and encrypted when you opt in.

Compact scenes

Compressed splat formats (SOG/SPZ) pack a building into tens of megabytes with no visible loss, small enough to text.

Where it's heading

The scan becomes the set.

Relight

Lighting the real room after the fact. Move the sun, change the hour, shape practicals on the scan you filmed. Blocking today; mood scouting next. In development.

Toward final pixels

The render path: taking the same scene from previs fidelity toward client-facing imagery. One file from scout to pitch.

Wider intake

More sources and formats over time. The bar is simple: if you can shoot it or build it, Zero should stand in it.

See one built, start to finish.

Get the beta

Intake

The quiet half.

Intake is the darkroom: footage goes in, a place comes out. It runs entirely on your machine, asks for minutes of video, and hands the result to Zero, or to anyone, as a single file.

IMPORTPOSE TRAINREFINE FINISH

The reconstruction stack

What actually happens in there.

Intake runs a full photogrammetry-class pipeline and a neural-rendering trainer on your own GPU. It is the stack research labs run on clusters, tuned to run beside your inbox.

Frame selection

The video is decimated to roughly 250-300 sharp frames. Motion blur and near-duplicates are dropped before anything expensive runs.

Pose solve

Structure-from-motion recovers every camera position to sub-pixel reprojection error, the same math survey-grade photogrammetry uses, minus the rig.

Gaussian optimization

Millions of anisotropic 3D gaussians are optimized on your GPU until the scene re-renders your own footage. Color, softness, and view-dependence are learned straight from the pixels.

Refine

You level the floor, crop the noise, and set the bounds, live, on the real data, exactly as it will ship.

Package

The scene compresses to SOG/SPZ, a building in tens of megabytes, and opens in Zero, or anywhere you send it.

~300 FRAMES · SUB-PIXEL POSE · MILLIONS OF GAUSSIANS · UNDER AN HOUR · 84 MB OUT

Shooting for the scan

What good footage looks like.

Move slowly

Walk the space like you're showing it to someone on a call. Smooth, unhurried, no whip pans.

Overlap everything

Each part of the place should appear from several angles. Circle objects that matter; cross the room more than once.

A few minutes is plenty

Two to five minutes of steady coverage, roughly 300 useful frames, beats twenty minutes of wandering.

Light it like you found it

Shoot the place as it is. The scan keeps the light you filmed. That honesty is the product.

Requirements

What it needs from your machine.

Building

A machine with a modern GPU, on Windows or Mac. Everything builds locally, and a typical scene finishes in under an hour.

Directing

Zero runs on far more modest hardware, and scenes travel: build on the workstation, walk it on the laptop.

Formats

Scenes save as compact splat files (SOG/SPZ/PLY); Zero also imports GLB and glTF for models and previs assets.

Footage in. A place out.

Get the beta

Pricing

One plan. Everything.

No tiers, no seats math, no feature ransom. The beta is free; this is what launch will look like.

At launch

$15 / month

7-DAY TRIAL · NO CARD · CANCEL ANYTIME

Zero + Intake, together
Unlimited scenes, built locally
Encrypted rooms for your whole team
Stills, plots, and MP4 export
Every update, one tier gets everything

Join the free beta

Beta users keep launch pricing.

Not unless you say so. Scenes build and live locally. If you open a room, the scene is end-to-end encrypted before it moves, and we relay ciphertext we cannot read.

Building in Intake wants a machine with a modern GPU, on Windows or Mac. Directing in Zero runs on much lighter machines.

You do. Entirely. They're files on your drive. We never see them, so there's nothing for a terms page to claim.

Yes. Rooms open from a link plus a password you share separately. No signup wall between your team and the set.

Free while it's in beta.

Get the beta

Early access

Get the beta.

Rolling invites, storytellers first. Signups open shortly. If you'd like to be in the first group, write to hello@takezero.io.

Early access

Arriving soon.

Invites open in the next few weeks. Want to be first? hello@takezero.io

While you wait

What you'll need.

To build

Windows or Mac, with a modern GPU.

To direct

Most modern machines. Scenes are small; nothing streams.

To start

A few minutes of steady video of a place you care about.

Legal

Privacy policy.

Effective July 21, 2026. Take Zero and Intake are made by Take Hold LLC. The short version: your footage, scans, and scenes never reach us. This page is mostly about the one thing you might give us: an email address.

What we collect

If you request beta access: your email address and what you shoot. Our hosting infrastructure also keeps standard server logs (IP address, pages requested, timestamps) for security and reliability.

What we never collect

Your video, scans, scenes, cameras, paths, and cuts. Builds run entirely on your machine. Collaboration rooms are end-to-end encrypted. The relay carries ciphertext we cannot decrypt, and room keys never leave your team.

Tracking

No advertising trackers and no cross-site profiles on this site. If that ever changes, this page changes first.

Your email

Used for beta invites and occasional product news. Every email includes an unsubscribe link. We never sell or rent your address.

Processors

The providers that host this site and deliver our email act as processors on our behalf, under their own data-protection commitments.

Your choices

Write to hello@takezero.io to see, correct, or delete anything we hold about you.

Changes

Updates are posted here with a new effective date.

Contact

Take Hold LLC · hello@takezero.io

Legal

Terms of service.

Effective July 21, 2026. These terms govern the Take Zero and Intake beta, operated by Take Hold LLC. Using the software means you agree to them.

Beta software

This is pre-release software provided as-is. Features may change, break, or be removed without notice. Keep your own copies of anything you care about.

Your content

Your footage, scans, and scenes are yours. We claim no ownership and take no license beyond what is technically required to run the features you invoke, for example relaying encrypted room data your team exchanges.

Capture responsibly

You are responsible for having the right to film the places and people you scan, and for complying with the laws that apply where you shoot.

Acceptable use

Don't use the software unlawfully, attempt to break or overload the service, or resell beta access. Beta invites are personal.

Access

We may modify, suspend, or revoke beta access at any time while the product is pre-release.

Pricing

The beta is free. Paid plans arrive at launch, with advance notice. Beta users keep launch pricing.

No warranty

To the fullest extent permitted by law, the software is provided "as is," without warranties of any kind, express or implied.

Limitation of liability

To the fullest extent permitted by law, Take Hold LLC is not liable for indirect, incidental, or consequential damages, and our total liability is limited to the amount you have paid us for the software.

Changes

We may update these terms; continued use after an update means you accept the new version.

Contact

Take Hold LLC · hello@takezero.io