API

Base URL https://api.studiorouter.com. For agents and LLMs: llms.txt, full agent guide, OpenAPI spec, MCP server at https://api.studiorouter.com/mcp.

1. Get a key

Your API key is in the dashboard; every account has one. Send it on every request as Authorization: Bearer sr_live_….

2. Start a generation

curl https://api.studiorouter.com/v1/videos \
  -H "Authorization: Bearer $STUDIOROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "model": "wan-3.0",
    "prompt": "A red kite over a quiet beach at sunset",
    "duration": 5,
    "resolution": "480p"
  }'

The answer is a job with status: "queued" and an estimated_cost_usd. Credits for the estimate are held while it runs. Sending the same Idempotency-Key again returns the same job instead of starting another.

3. Wait for the result

curl "https://api.studiorouter.com/v1/jobs/job_…?wait=50" \
  -H "Authorization: Bearer $STUDIOROUTER_API_KEY"

wait holds the request open for up to 50 seconds until the job finishes. Repeat until status is succeeded, failed or blocked. On success output[0].url is the file, hosted by us, and cost_usd is what was charged: the provider's own price. Failed and blocked jobs are not charged.

Video parameters

The input mode follows from what you send: nothing extra is text-to-video, image is first-frame, image + last_image is first-and-last frame, any reference_* list is reference mode, and video is an edit.

Images

POST /v1/images takes model, prompt, optional images (input URLs, which makes it an edit), resolution (1k, 2k, 4k), aspect_ratio, quality (GPT Image models), output_format and seed where supported.

Local files

Inputs are public URLs. To use a file from disk, upload it first and use the returned URL:

curl https://api.studiorouter.com/v1/files -H "Authorization: Bearer $STUDIOROUTER_API_KEY" \
  -H "Content-Type: image/png" --data-binary @first-frame.png

Price a request first

POST /v1/estimate takes the same body as a generation and returns the estimated cost for each provider that can serve it, cheapest first. Nothing is created.

Agents (MCP)

Step-by-step setup for Claude Code, Codex and Claude Desktop is on Connect an agent.

claude mcp add --transport http studiorouter https://api.studiorouter.com/mcp \
  --header "Authorization: Bearer sr_live_…"

Tools: list_models, get_model, estimate_cost, generate, get_job, get_balance. generate and get_job accept wait_seconds (up to 50) so an agent can wait for the result without polling in a loop.

Errors

Errors are JSON with a message that says what went wrong. Failed and blocked jobs are not charged.

Model ids

Prices and providers for each model are on its page under Models, and in GET /v1/models.