Measured cost × 3. Quoted before work.
No subscription. USDC on Base through x402. Each request's price is three times its measured variable cost (minimum 0.010 USDC), computed from the cost drivers you choose and bound into the payment challenge before any extraction starts.
| Request shape | Price (USDC) | Expected variable cost |
|---|---|---|
| Deterministic only (strategy=deterministic) | 0.010 | $0.0033 |
| Single record (default) | 0.031 | $0.0101 |
| List up to 10 items | 0.070 | $0.0232 |
| List up to 25 items | 0.140 | $0.0464 |
| List up to 50 items | 0.255 | $0.0849 |
| Single record + 1 followed link | 0.044 | $0.0146 |
| Single record + 2 followed links | 0.058 | $0.0191 |
| Single record with JavaScript rendering | 0.036 | $0.0119 |
What drives the price
- Model use. strategy=deterministic never calls a model and is the cheapest. strategy=auto reserves one semantic extraction call.
- Output size. Single-record requests are cheapest. Setting max_items or an array schema selects a list budget (≤10, ≤25, ≤50 items).
- Pages. Each followed link (follow_links 1–2) adds retrieval and model input.
- Rendering. render_javascript=always adds a rendering reserve. With auto, JavaScript shells are rendered when needed at no extra charge.
Cost basis
Rates (USD): model input $1/M tokens, output $5/M tokens; payment facilitator $0.0010; database $0.0022; compute 0.0000415/s; rendering $0.0015/page. Token expectations come from real-provider acceptance runs measured on 2026-09-28; see the docs for details.
Get the exact price for free with POST /api/v1/quote, or call /api/v1/extract and read the 402 response.
Price = 3 × measured variable cost for the request's cost drivers (model use, output size, pages, rendering), quoted before work and bound into the x402 challenge. Valid partial/empty results are chargeable. Provider failures allow two free retries with the same Idempotency-Key; no automatic refund. Uncertain settlement requires reconciliation.