Limits
Request validation, retention, and plan capacity limits.
Request and retention limits
- Data API list endpoints default to 50 items per page and accept a
limitof 1–100. - Extraction fields: 200 per request; extraction name: 160 characters.
- Extraction description: 1,000 characters.
- Extraction instructions: 12,000 characters.
- Browser actions: 20 per Data API request; workflow definition steps: 50.
- Recipe network JSON capture: up to 8 unique endpoints, 256 KiB per response, and 1 MiB total during the first 10 seconds or the original page-request deadline, whichever comes first.
- Cache TTL: 0–3,600 seconds.
- Data API timeout: 5,000–300,000 ms; default 45,000 ms.
- Data API result retention: 7 days; trace, usage, and audit retention: 90 days.
Plan capacity
| Plan | API keys | Concurrent jobs | Batch | Requests/minute | Effective Data API timeout | External CAPTCHA attempts |
|---|---|---|---|---|---|---|
| Developer | 10 | 10 | 250 | 300 | 120s | 3 |
| Business | 50 | 50 | 1,000 | 1,200 | 300s | 5 |
| Scale | 100 | 100 | 5,000 | 3,000 | 300s | 5 |
| Advanced | 250 | 250 | 10,000 | 6,000 | 300s | 8 |
The effective timeout is the smaller of the plan allowance and the Data API's absolute 300-second validation limit. A higher account allowance does not let a request exceed that schema limit.
The account response and server validation are authoritative for the current plan. Keep these numbers out of client-side enforcement logic; use them for capacity planning and friendly preflight messages.
Last updated on