Skip to main content
Save 20% when you pay yearlySee plans

Use case: Developer automation

Make images and videos from your own code or your AI assistant

Send a prompt over HTTPS, or add Create to Claude or ChatGPT and just ask. The same models and the same credits as the app, with a daily cap on every key.

curl
curl https://create.vin/api/v1/images \
  -H "Authorization: Bearer crk_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "A lighthouse at dusk, soft film look", "aspect": "16:9"}'

# {
#   "id": "...",
#   "status": "complete",
#   "output_url": "https://...",
#   "credits_charged": 12
# }

The numbers in the answer are an example. Real prices are in GET /api/v1/models.

What you can automate

One key works for the REST API and the MCP server.

Images from a script

POST a prompt and, if you like, an aspect ratio or a model. The call waits and answers with a link to the picture.

Videos from a script

POST a prompt, a length and an optional starting picture. You get an id straight away, then check on it until the video is ready.

Ask your assistant instead

Add Create to Claude, ChatGPT or another MCP app, then ask for a picture or a clip in plain words and get the link back.

Spending you control

Each key has its own daily credit cap, 30 requests a minute, and you can revoke it at any time. A generation that fails is refunded.

Using a code generator or an API client? Download the OpenAPI description (/api/v1/openapi.json).

How it works

From a paid account to a first result in three steps.

  1. 1

    Make a key

    Sign in, open the API page and make a key. It starts with crk_live_ and is shown once.

  2. 2

    Send it as a Bearer token

    Put it in the Authorization header of every API call, or give it to your assistant app as a header for https://create.vin/api/mcp.

  3. 3

    Call a route or ask

    Post to /api/v1/images or /api/v1/videos, or ask your assistant for a picture. The examples below show both ways.

Make an image

The call waits for the picture and answers with a link. The link works for about an hour, so save the file if you want to keep it.

curl
curl https://create.vin/api/v1/images \
  -H "Authorization: Bearer crk_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "A lighthouse at dusk, soft film look", "aspect": "16:9"}'

Make a video, then wait for it

A video takes a minute or more, so the call answers at once with an id and status 202. Ask for that id every few seconds until the status is complete or failed.

curl
# 1. Start the video
curl https://create.vin/api/v1/videos \
  -H "Authorization: Bearer crk_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Waves rolling onto a beach", "duration": 5, "aspect": "16:9"}'

# 2. Check on it (use the id from step 1)
curl https://create.vin/api/v1/generations/GENERATION_ID \
  -H "Authorization: Bearer crk_live_YOUR_KEY"

Connect an assistant

Five tools show up in your assistant. You do not call them yourself, you just ask.

Claude Code
claude mcp add --transport http create https://create.vin/api/mcp \
  --header "Authorization: Bearer crk_live_YOUR_KEY"
ToolCreditsWhat it does
create_imageSpends creditsMakes a picture from a text prompt and returns a link to it. Inputs: prompt, aspect, model.
create_videoSpends creditsStarts a video from a prompt, optionally from a starting picture, and returns an id. Inputs: prompt, image_url, duration, aspect, model.
get_generationNo creditsLooks up one picture or video by id: pending, complete or failed, and the link when it is ready. Input: id.
list_my_generationsNo creditsLists the 20 most recent things made on the account, newest first, with links.
get_creditsNo creditsShows the credits left on the account and this key's daily limit.

Steps for Claude Desktop, ChatGPT and other apps are on the MCP server page.

Plans needed

The API and the MCP server are for paid plans. Any paid plan works, and its credits are the ones you spend.

Plans start at Basic, $12 a month with 1,000 credits. There is no separate API balance and no extra API fee: each image or video takes its price in credits from your account, the same price as in the app.

See plans and credits

Limits, stated plainly

Worth knowing before you build on it.

  • Paid plans only. There is no way to use a key without one.
  • 30 requests a minute for each key. Every call counts, reads included.
  • A daily credit cap on every key: 500 to start, up to 5,000, reset at 00:00 UTC.
  • Up to 200 credits for one MCP call.
  • Up to 5 active keys per account. A key is shown once when you make it, so copy it right away.
  • Pictures and videos only. Links to finished files work for about 60 minutes, so save the file if you want to keep it.
  • An app that can only sign in through a pop-up window, and cannot send an Authorization header, cannot connect to the MCP server.

Over the request limit you get a 429 with a Retry-After header.

Get your API key

Sign in, make a key, set its daily cap and start calling. You can revoke it whenever you want.

Questions