Limits
These values were read from the running service (GET /api/v1/info/status) and are refreshed hourly.
| Limit | Value | Notes |
|---|---|---|
| File size on the website | 500 MB | Per file. Larger files are refused before upload. |
| Sent in one piece | up to 95 MB | Above this the website switches to the chunked upload API on its own. |
| API request body | 100 MB | One direct request. Use /api/v1/uploads for anything larger. |
| Chunked upload | 500 MB per file, 32 MB chunks | The SDK and CLI use it automatically for large inputs. |
| Batch in the browser | 20 files, 1024 MB in total | Single-file tools run them one by one with the same settings. |
| OCR | 50 pages per request that need recognition | Pages that already contain text are skipped and do not count. |
MB here is 1,048,576 bytes, the unit the tools use. A single tool may accept fewer file types than any file: its page lists the formats. See the developer docs for the upload API and error codes for what a request over a limit returns.
OCR languages
The Tesseract language packs installed on the service (20): English, German, French, Spanish, Italian, Portuguese, Dutch, Russian, Ukrainian, Polish, Turkish, Arabic, Hindi, Simplified Chinese, Traditional Chinese, Japanese, Korean, Vietnamese, Thai, Indonesian. Up to four can be combined in one request.
Rate-limit headers
API responses carry X-RateLimit-Limit, X-RateLimit-Remaining and X-RateLimit-Reset. The website and API work without an account.