Document Processing Overview
Parselyze exposes two public document-processing modes:
- Synchronous: parse now, get JSON in the same response
- Asynchronous: submit a job now, fetch results later or receive them by webhook
Both modes use the same template-based extraction engine.
Which One Should You Choose?
| Use case | Recommended mode |
|---|---|
| Immediate request/response flow | Synchronous |
| Background workers or queues | Asynchronous |
| Webhook-driven automation | Asynchronous |
| Multiple uploaded files in one request | Synchronous |
| ZIP archives | Synchronous |
| One file per background job | Asynchronous |
Sync vs Async
| Feature | Synchronous | Asynchronous |
|---|---|---|
| Response style | JSON returned immediately | Job returned immediately, result later |
| Upload shape | files multipart field | file multipart field |
| Multiple files | Yes, up to 10 uploaded files | No |
| ZIP support | Yes | No |
| Webhooks | No | Yes |
| Polling | Not needed | Yes, via GET /v1/jobs/:jobId |
| Best for | Simple parse requests | Background and automation workflows |
Shared Behavior
Both modes:
- Require an existing template ID
- Support PDFs and common image formats
- Accept an optional
languagehint - Support
fileUrlas an alternative to direct upload
Current Product Limits
- Templates are created in the dashboard, not through the public API
- Async processing supports one file per job
- Scanned PDFs over 30 pages are not supported yet
Public Endpoints
- Synchronous:
POST /documents/parse - Asynchronous submit:
POST /v1/documents/parse/async - Async job status:
GET /v1/jobs/:jobId