Skip to content
Get started

API

Drive it from your own code

Everything the app does, over HTTP: images, video, avatars, voiceovers, music and carousels. One token, one header, no SDK to install. 45 endpoints, 17 of them free.

Start in two minutes

  1. 1. Create a token

    In Settings → API tokens, name it after the machine it will run on and copy it. It looks like ait_… and is shown once. It acts as your account and spends your credits, so treat it like a password — you can revoke it from the same page at any time.

  2. 2. Check it works

    This call is free and returns your balance, your plan and your projects. Most endpoints need a projectId, and this is where you get one.

    curl https://aitachyon.com/api/me \
  -H "Authorization: Bearer ait_your_token_here"

  3. 3. Generate something

    curl https://aitachyon.com/api/images/generate \
  -H "Authorization: Bearer ait_your_token_here" \
  -H "Content-Type: application/json" \
  -d '{"projectId":"<your project id>","prompt":"a flat lay of our product on linen"}'

Everything bills in credits. A call that costs credits spends them for real — there is no sandbox. Free endpoints are marked below, and they cover everything you need to explore before committing.

Video and avatar calls are asynchronous. They return a task id; poll the matching status endpoint until the URL appears. Those are marked too.

Failures are refunded. A job that fails returns its credits, and the refund is linked to the job in /api/credits/ledger.

Endpoints

Account

Start here. Free, and the fastest way to check a token works.

GET/api/mefree

Your credit balance, plan, projects and available avatars.

GET/api/credits/ledgerfree

Every spend and refund, each linked to the job that caused it.

GET/api/generationsfree

What you have generated, newest first.

Images

Text-to-image and image-to-image, on-brand when a brand is attached.

POST/api/images/generate2 cr · 1-4 depending on provider, +1 for the skin finish

Generate an image. Providers differ in price; a people shot may add a finishing pass.

POST/api/images/edit2 cr

Remove background, upscale, vectorize, replace background or inpaint.

POST/api/images/director-previewfree

See the art-directed prompt before spending anything.

POST/api/images/ideasfree

Image concepts for a brand or product.

Video

Scene generation and post-production. These are asynchronous: you get a task id to poll.

POST/api/scenes/generateasync3 cr · 3 fast, 28 hd, or per second by engine

Generate a b-roll clip. Eight engines; flat-priced on the default tier, per-second on the others.

GET/api/scenes/status/[taskId]free

Poll a scene until the mp4 is ready.

POST/api/scenes/director-previewfree

Preview the art-directed video prompt, free.

POST/api/videos/upscaleasync1 cr · per second of output

Upscale a finished video to 1080p or 4K.

POST/api/videos/reframeasync5 cr · per second

Reframe to another aspect ratio, generatively.

POST/api/videos/motionasync4 cr · per second

Transfer the motion and camera of a driving clip onto a still.

POST/api/videos/lipsyncasync4 cr · per second

Retrofit lipsync onto an existing clip.

POST/api/videos/dubasync18 cr · per minute, rounded up

Dub a video into another language.

POST/api/videos/matteasync1 cr

Remove the background, green-screen output.

POST/api/videos/analyze2 cr

Score a finished video for virality before you post it.

POST/api/export/assembleasync1 cr

Assemble clips, voiceover and captions into a finished mp4.

Avatars and presenters

Talking-head video from a photo or a trained likeness.

POST/api/avatar/generateasync5 cr · per second, minimum 30

Avatar video from an avatar and a voiceover.

GET/api/avatar/status/[predictionId]free

Poll an avatar render.

POST/api/avatars/persona2 cr

Create a reusable persona from a description.

POST/api/avatars/twin2 cr · +28 to train a LoRA

Build a likeness of yourself from reference photos.

POST/api/creator/presenter2 cr

A presenter shot matched to your brand.

POST/api/creator/planfree

Plan a creator-style shoot, free.

Audio

Voiceover, cloned voices, music and sound effects.

POST/api/voiceover/generate1 cr · 1-5 per 450 characters by provider

Voiceover from a script.

GET/api/voiceover/voicesfree

Every voice you can use.

POST/api/voiceover/samplefree

Hear a voice before committing, free.

POST/api/voices/clone20 cr

Clone a voice from a sample; returns a reusable voice id.

POST/api/voices/speak3 cr · per 450 characters

Speak with a cloned voice.

POST/api/music/generate10 cr

A background track.

POST/api/sfx/generate1 cr

A short sound effect.

Brand and copy

Read a website once, then everything you generate pulls from it.

POST/api/brand/reference-kitfree

Read a URL for its logo, colours, fonts and usable imagery. Free, and nothing is created.

POST/api/brand/scrape1 cr

Analyse a website into a brand kit.

POST/api/brand/create1 cr

Create a brand kit by hand.

POST/api/brand/logo2 cr

Generate a logo.

POST/api/scripts/generate2 cr

An ad script from a brief.

POST/api/copyfree

Headlines, CTAs and captions, free.

POST/api/content/generate-captions1 cr

Platform-sized captions.

POST/api/campaigns/angles2 cr

Distinct campaign angles for one product.

POST/api/design-carousel/generate2 cr

A code-rendered carousel on your real brand theme.

3D and assets

Models, uploads and signed URLs.

POST/api/three-d/generateasync10 cr

An image turned into a downloadable GLB.

GET/api/three-d/status/[jobId]free

Poll a 3D job.

POST/api/uploadfree

Upload an image, video or audio file to your account.

POST/api/signed-urlfree

A signed URL for one of your assets.

POST/api/inspiration/searchfree

Search reference imagery, free.

Prefer to drive it from Claude Code or Cursor? The same account works over MCP — create a token and the setup block is right there.