Document Processing Overview

Synchronous Processing Deprecated

The synchronous processing method is deprecated and will be removed in a future version.

We strongly recommend using Asynchronous Processing with Webhooks for:

• Better reliability with automatic retries
• No timeout limitations
• Idempotent job submission
• Ability to retrieve results multiple times

The async method is now the recommended approach for all use cases.

Parselyze offers two ways to process documents: Synchronous (deprecated) and Asynchronous (recommended). Choose the approach that best fits your use case and workflow requirements.

Sync vs Async: Which to Choose?

Feature	Synchronous (Deprecated)	Asynchronous (Recommended)
Response Time	Immediate (within seconds)	Job ID returned instantly, results later
Best For	Small documents (1-5 pages)	All documents (any size)
Timeout	30 seconds max	No timeout, processes in background
Integration	Simple HTTP request/response	Webhooks or polling
Reliability	Single attempt	3 retry attempts with backoff
Result Access	One-time response only	Retrievable for 7 days via job ID
Rate Limits	Same limits apply	Same limits apply
Idempotency	Not built-in	Automatic duplicate detection
Status	⚠️ Deprecated	✅ Recommended

⚠️ Synchronous (Deprecated - Not Recommended)

Deprecated

Synchronous processing is deprecated. Please migrate to asynchronous processing.

Use synchronous processing only if:

• Documents are small (under 5 pages)
• You need results immediately in the response
• Simple request/response pattern works for your use case

Legacy documentation for Synchronous Processing →

✅ Asynchronous (Recommended)

Use asynchronous processing for all use cases:

• ✅ All document sizes (small to large)
• ✅ Better reliability with automatic retries
• ✅ No timeout constraints
• ✅ Webhook notifications for real-time updates
• ✅ Idempotent job submission (duplicate prevention)
• ✅ Results accessible for 7 days
• ✅ Perfect for batch processing
• ✅ Integration with background job systems

Example: Invoice processing, receipt scanning, form extraction, bulk document analysis, automated workflows

Learn more about Asynchronous Processing →

Quick Comparison

Synchronous Flow

Upload Document → Wait for Processing → Receive Results → Done

Total Time: 2-10 seconds (one request)

Asynchronous Flow

Submit Job → Receive Job ID → Continue Work → Webhook Notification → Retrieve Results → Done

Total Time: Instant submission + background processing

API Endpoints

• Synchronous: POST /documents/parse
• Asynchronous: POST /v1/documents/parse/async
• Job Status: GET /v1/jobs/:jobId

Next Steps

1. Authentication: Set up your API key
2. Synchronous Processing: Process small documents immediately
3. Asynchronous Processing: Handle large documents with retries
4. Webhooks: Receive automatic notifications
5. Error Handling: Handle errors gracefully