Skip to content
NewOur own model: about $0.03 a second

Describe a video.
We make the whole thing.

One agent researches, scripts, narrates, shoots and cuts it. Ten seconds to two hours, on any subject.

Our model≈$0.03 / s

One line in, one finished video out4.9 from 7,000+ creatorsMillions of YouTube views

Documentary · 16:9
Rendering

What people make with it

Built on frontier models from

Anthropic
OpenAI
Gemini
Grok
MiniMax
Seedance
Anthropic
OpenAI
Gemini
Grok
MiniMax
Seedance
Anthropic
OpenAI
Gemini
Grok
MiniMax
Seedance

Made with NoodleTomato

Real exports, our model, and concept renders. Click any clip for the prompt.

Styles

Every style the agent can shoot in, from documentary to claymation. Describe your own if none fits.

Illustrative shots, each generated from the prompt shown. For finished videos and the prompts behind them, see the examples above.

NoodleTomato Motion

Our own model

Try the model

Our own model. About $0.03 a second, audio included.

Skip the agent, write the prompt yourself, and NoodleTomato Motion renders the clip: picture and sound in one pass, from text or a reference image. We run it on our own hardware, so a generated second costs $0.0335 in landscape, $0.023 in portrait and $0.0425 in square. A ten-second widescreen clip is about $0.34. The third-party engine in the same studio is $0.115 a second. You see the exact price before anything renders. Six clips below, straight out of it.

See every rate
Nature: bioluminescent waves glowing blue on a night beach, a distant lighthouse beam, stars, gentle surf
Nature: a macro bee landing on a lavender flower, pollen dust, morning light, soft buzz
Nature: a red desert canyon river at sunrise from a drone, long shadows, slow forward flight, wind
Nature vertical: a waterfall in a green gorge with mist and a rainbow, slow tilt down, roar
Nature: a chameleon eye in extreme macro, scales catching light, slow turn, jungle ambience
History vertical: a knight's gauntlet gripping a sword hilt, banners moving in wind, morning fog
Text or reference imagePicture and sound in one pass768p · 24 fps4 to 15 seconds a clip$0.0335 a second in landscape$0.023 portrait · $0.0425 squareExact quote before it renders
Try the model

Unedited samples.

You write one line. It runs the production.

The agent works in stages and shows you each one.

Surviving alone · 16:9

Surviving alone

Create a 30-minute YouTube documentary about what it really takes to survive alone in the wild.

Shot 2 · The tree line
Shot 6 · First fire
Shot 11 · Shelter
Shot 15 · Water
AgentReading 14 sources

Got it. I'll research the subject and come back with a brief before I write a line.

NARRATOR

TENS OF THOUSANDS OF VOICES
ONE LINE OF SCRIPT

Narrators and dialogue voices. Pick one per video, or none.

Documentary
  • Documentary

    Male · Deep · Measured

  • Alice

    Female · Clear · UK

  • Bill

    Male · Wise · US

Conversational
  • Warm

    Female · Soft · Conversational

  • Chris

    Male · Casual · US

  • Jessica

    Female · Playful · US

Cinematic
  • Cinematic

    Male · British · Grave

  • Lily

    Female · Velvety · UK

  • Adam

    Male · Firm · US

Bright
  • Bright

    Female · Clear · Upbeat

  • Liam

    Male · Energetic · US

  • Bella

    Female · Bright · US

Storyteller
  • Storyteller

    Male · Warm · Intimate

  • Claire

    Female · Cozy · US

  • George

    Male · Storyteller · UK

Character
  • Harry

    Male · Fierce · US

  • Callum

    Male · Husky · US

  • Laura

    Female · Quirky · US

Eighteen samples reading one line of the script. The studio holds tens of thousands of voices, narrators and dialogue alike.

It remembers you.

Your narrator, your look, your characters and every episode so far stay on file. Come back next month and it picks up mid-sentence.

Surviving alone · Episode 2

Session 2 · three weeks after episode 1

You

Episode 2. Same series, same everything.

Memory updated · Surviving alone is a series
AgentReading 1 project · 14 notes

Picking up Surviving alone where episode 1 left off. Brian narrates, same handheld look, 30 minutes in 16:9, no music under narration. Dan keeps the red jacket and the beanie. Episode 1 ended at 29:41 with the helicopter passing over the valley, so this one opens on the sound of it fading. Writing the brief now.

From memory

  • Narrator: Brian
  • Handheld documentary
  • Dan's red jacket
  • Opens where 29:41 ended
  • Characters stay themselves

    Same faces, same outfits, same voices, episode after episode.

  • Ask it anything about the record

    What happened in episode four, which take you rejected, why. It answers from the file, with the timestamp.

  • Said once, kept

    Narrator, look, format, pacing and the things you always fix carry into every new project.

Long videos, made without prompting, for very little.

Any style, any subject

Fifteen styles on the page, and you can describe your own.

  • Runtime

    Ten seconds to two hours

    It plans the whole runtime before it renders a frame.

  • Control

    You direct. It revises.

    Say what to change and it changes that part.

  • Output

    Any format

    Widescreen, vertical or square from the same run.

  • Voice

    Pick the voice

    Tens of thousands of narrators and dialogue voices, one per video.

  • It remembers you

    Narrator, look, characters and every episode stay on file between sessions, so the next one starts where the last ended.

    Memory
  • It reads before it writes

    Sources first, then the brief, then the script. Nothing in the narration is made up on the spot.

    Research
  • No AI video to render

    The agent can work with real footage instead of generating it. Any reasoning, voice, music or generated images you use still consume credits.

    Real footage

A video agent for any kind of video.

A documentary, a product ad, a micro drama, a course. Describe it once; the agent researches, scripts, voices and cuts it, and you review each stage.

NoodleTomato
Your agent
Codex
MCP
API

BRING YOUR OWN AGENTNEW

Connect your compatible agent client over MCP. Your agent does the thinking on the plan you already pay for; NoodleTomato makes the images, video, voice and music and charges credits for those.

Connect your agent

Needs an active NoodleTomato plan. Your reasoning is billed by your own client, never by us; media comes out of your NoodleTomato credits.

The same engine, over an API.

Text to video, image to video and reference images from your own code, with the same quotes, credits and ledger as the studio. Any active plan includes it.

Catalog

Options and capabilities

List the generation options, sizes, durations and image quality levels your key can use, then request a quote for your chosen settings.

Text to video

A prompt in, a clip out

Send a prompt and get a finished clip. One quote covers the whole job, reference frame included.

Image to video

Animate your own frame

Start from an image you upload or one you made here, pick length and size, and get the clip back.

Reference images

Frames for your shots

Make reference images at an explicit quality level, and reuse them across clips for a consistent look.

Async jobs

Submit, poll, cancel, fetch

Every generation is a job: submit it, check its status, cancel it while it is queued, and download the assets when it is done.

Quotes and usage

Know the price first

Ask for a quote before you submit, see your balance and usage, and reserve credits when you submit a job.

Auth

Scoped, revocable keys

Create keys in your developer settings, scope them to what they may do, and revoke them any time. Keys need an active plan.

Get a key

Keys live in your developer settings. The OpenAPI document describes every endpoint.

NoodleTomato Motion is $0.0335 per generated second in landscape, NoodleTomato Motion HD from $0.0625, and a reference frame from $0.0041. The API quotes in credits, at 100 credits to the dollar. Credits are reserved on submission. Final charges stay within the quote; unused reserved credits are released. Price list, quick start and endpoints at noodletomato.com/developers.

Pay for what you use.

A monthly plan opens the studio, and what you pay comes back as usage: 100 credits = $1 of it. Every job shows its price before it runs, and you only spend what the work costs.

Starter

Try the agent, our own model and the voice library on a small monthly budget.

  • Usage: $9 a month, as 900 credits
  • Renders: about 4 min of Motion footage, sound included
  • Access: agent, both models, voices, MCP and API
$9per month
Most popular

Creator

For one channel or one series that ships every week.

  • Usage: $29 a month, as 2,900 credits
  • Renders: about 14 min of Motion footage, sound included
  • Access: agent, both models, voices, MCP and API
$29per month

Studio

For repeated production: several projects, longer runtimes, more revisions.

  • Usage: $79 a month, as 7,900 credits
  • Renders: about 39 min of Motion footage, sound included
  • Access: agent, both models, voices, MCP and API
$79per month

Scale

For programmatic output over the API and MCP, or several channels at once.

  • Usage: $199 a month, as 19,900 credits
  • Renders: about 99 min of Motion footage, sound included
  • Access: agent, both models, voices, MCP and API
$199per month

Footage figures are landscape clips on NoodleTomato Motion at $0.0335 a generated second, from your own starting images. Reference frames, the agent, voices and music are extra; the prices are below.

What things cost

Prices in US dollars. Plans and top-ups buy credits at 100 credits to the dollar, so a job that costs $1.00 spends 100 credits. The same prices apply in the studio, over MCP and over the API, on every plan.

Video engines, price in US dollars per generated second
Video, per generated second16:99:161:1
NoodleTomato MotionOur own model$0.0335$0.023$0.0425
Picture and sound in one pass, from text or a starting image. 4 to 15 seconds a clip.
NoodleTomato Motion HDOur own model$0.133$0.1455$0.0625
High-resolution clips from text, a starting frame or visual references. 4 to 12 seconds a clip.
CinematicThird-party$0.115$0.115$0.115
A third-party engine offered in the same studio, at 720p. 4 to 15 seconds a clip.

Text to video on our own models makes one reference frame per clip first, charged as an image below. A starting image you supply adds no image charge.

Images, per image, landscape or portrait

  • Detail imageFive quality levels, from quick drafts to maximum fidelity. The better the reference frame, the better the clip.
    Low$0.0041Medium$0.0097High$0.0372Extra high$0.0661Maximum$0.1488
  • ImageThe standard still-image option, three quality levels.
    Low$0.0058Medium$0.0472High$0.1898

Voice, music and the agent

  • NarrationAbout $0.34 for a minute of speech.$0.4888 per 1,000 characters
  • Music bedAn instrumental bed up to three minutes long.$0.069 per track
  • Agent reasoningCharged by how much a step reads and writes. No charge from us when you bring your own agent over MCP.$0.0007 per step, typical

Video prices are per generated second, sound included. Every job shows its price before it runs and never costs more than that. A retake or a revision is a new generation.

Three real receipts

Jobs from our own production account, September 2026, as they were charged.

  • A 4-second clip, text to videoNoodleTomato Motion · landscape · with sound
    • Video clip, 4 s$0.134
    • Reference frame, low quality$0.0045

    Charged$0.1385 13.85 credits

  • A 3.75-second clip in high resolutionNoodleTomato Motion HD · 2688 × 1536 · with sound
    • Video clip, 3.75 s$0.4987
    • Reference frame, low quality$0.0045

    Charged$0.5032 50.32 credits

  • One step of agent reasoningDeep reasoning mode · one response
    • Reasoning, one step$0.0007

    Charged$0.0007 0.07 credits

Work out a video

Pick the engine and the runtime and see what it comes to, and which plan covers it.

Engine
Format
10 s1 hour
Add

Estimate

$2.11210.7 credits

  • NoodleTomato Motion, 60 s$2.01
  • 10 reference frames, medium quality$0.10
  • NoodleTomato Motion, landscape$0.0335 a second

Covered by Starter at $9 a month.

A planning estimate for first takes: six-second clips, medium reference frames. The agent's reasoning adds a little, and retakes or revisions are new generations. The studio shows the exact quote before each job.

Top-ups

Add credits to any active plan, the $9 one included, at the same rate. Bought and spent inside the app.

  • $10· 1,000 credits
  • $25· 2,500 credits
  • $50· 5,000 credits
  • $100· 10,000 credits
  • $250· 25,000 credits
  • $500· 50,000 credits
  • $1,000· 100,000 credits
  • $2,500· 250,000 credits
  • $5,000· 500,000 credits

Every plan includes

  • The agent, the studio, NoodleTomato Motion and Motion HD
  • Model and quality choices on every job
  • The voice library and project memory
  • Bring your own agent over MCP
  • API access with scoped keys

How credits work

  • 100 credits are $1 of usage, on every plan and every top-up.
  • Every job is quoted before it runs. Video is charged per generated second of the option you pick, images per image, voice and music per use, and NoodleTomato-hosted reasoning per run.
  • Credits from a plan or a top-up expire 2 calendar months after they are issued. A renewal or a top-up adds a new batch with its own expiry and does not extend older credits.
  • An active plan, any tier, is required to use the studio, the agent, MCP and the API, and to buy or spend top-ups. Cancel any time; access runs to the end of the paid month.

Prices in USD, billed monthly. Bring your own agent over MCP and your reasoning is paid in your own client; NoodleTomato charges credits only for the media it makes.

Four ways it pays.

A finished video is an asset you own. YouTube is the most common place to put it to work, and it is not the only one.

  1. YouTube ad revenue

    Build a channel, publish on a schedule and earn a share of the ads through the Partner Program. AI-assisted video is eligible under the same rules as anything else, with disclosure where YouTube asks for it.

  2. Micro drama and serialized fiction

    Short scripted episodes with recurring characters, earning through the apps that carry them or ads against the episodes. The cast, the look and the narrator stay on file for the whole series.

  3. Video for clients

    Agencies, studios and freelancers deliver client work at a volume filming cannot reach, and pitch with a finished cut instead of a treatment.

  4. The audience you already have

    Publishers, newsletters and course sellers turn the writing they already do into video, and monetize it the way they monetize everything else.

We do not publish income estimates. What any of this makes depends on your subject and your audience, and anyone quoting you a monthly number is guessing. What it costs to make is the number we can print.

A river winding through a red rock canyon at golden hour

Niche Finder

For YouTube channels

Open the finder

Questions we get a lot

  • One finished video file. The narration, the generated visuals, the music, the subtitles and the edit are all already in it. Most people upload it as it comes.
  • You can direct it. It shows you the brief, the hook and the full script before it generates a single frame, and you can rewrite any of that or ask for another pass.
  • Yes. Every project keeps its style guide, narrator, characters, scripts and exports, and what you tell it once carries into the next project. Open it a month later and it picks up where you left off; ask what happened in episode four and it answers from the record.
  • Monthly plans from $9 to $199, with $1,000, $2,500 and $5,000 plans for high volume. What you pay comes back as usage, counted in credits, and one credit is one cent. NoodleTomato Motion, our own model, costs $0.0335 per generated second in landscape, $0.023 in portrait and $0.0425 in square, sound included: about $0.03 a second. NoodleTomato Motion HD starts at $0.0625 a second, the third-party engine is $0.115, a reference frame on the Detail image engine runs from $0.0041 to $0.1488 by quality, narration is about $0.34 a minute and a music bed is $0.069. The full price list is in the pricing section. Every job shows its price before it runs and never costs more than that. Working with real footage avoids AI-generated video costs. Any reasoning, voice, music or generated images you use are still charged.
  • Each batch of credits, from a plan or a top-up, expires two calendar months after it is issued. Renewing or topping up adds a new batch with its own expiry and does not reset older credits. Top-ups go from $10 to $5,000 at the same rate, and an active plan of any size is required to buy or spend them.
  • Yes. Connect a compatible MCP client from your developer settings in the app, where you will find setup instructions. Your agent does the reasoning in the client you already pay for, and NoodleTomato charges credits for the images, video, voice and music it makes. An active NoodleTomato plan is required.
  • Yes, on every plan. Text to video, image to video and reference images, plus quotes, asynchronous jobs with status and cancellation, asset download, balance and usage, behind scoped keys you create and revoke in your developer settings.
  • No. YouTube is where most of the output ends up, but the agent does not care where a video goes. People use it for serialized micro dramas, client work, course and newsletter content, and channels on other platforms.

Start with

The life of a Roman soldier
History · 30 minThe life of a Roman soldier
How the golden retriever won
Documentary · 10 minHow the golden retriever won
Is the American Dream still reachable?
Documentary · 20 minIs the American Dream still reachable?
What surviving alone actually takes
Documentary · 30 minWhat surviving alone actually takes
The cars that will redesign cities
Science and engineering · 30 minThe cars that will redesign cities
Noodle Tomato: AI Faceless YouTube Video Generator