API v1

API documentation

The same ranked search as the website, over HTTP. Create a key in your dashboard. Each account gets 100 free billable requests per UTC day, counted across all its keys. After that each billable request costs $0.01 from your credit balance; failed searches are refunded. Responses include free_requests_left_today, charged_usd and balance_usd (null when nothing was charged). Rate limit: 60 requests/minute per key.

POST /api/v1/search

Rank offerings in one market. One billable request.

curl -X POST https://YOUR_DOMAIN/api/v1/search \
  -H "Authorization: Bearer aac_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "category": "video-generation",
    "filters": [
      { "attr": "native_audio", "op": "has", "value": true },
      { "attr": "max_resolution", "op": "eq", "value": "1080p" }
    ],
    "maxPrice": 0.5,
    "weights": { "cost": 0.7, "quality": 0.3, "speed": 0 },
    "sort": "score",
    "limit": 5
  }'

Errors: 400 invalid_query, 401 bad key, 402 insufficient_credits (free allowance used up and no credit left), 429 rate limited.

GET /api/v1/offerings

Without ?category: list of markets and their attributes (free). With it: every offering and price in that field (one billable request).

curl "https://YOUR_DOMAIN/api/v1/offerings?category=gpu-on-demand" \
  -H "Authorization: Bearer aac_live_..."

Markets and attributes