PARSER GUIDES

Processing from URLs

Ways to Add Documents

Process documents from URLs via the API

PARSER can process documents from URLs--you send links to documents (e.g. PDFs on a server or cloud storage) instead of uploading files. The system downloads each file from the URL and processes it like a normal upload.

This is done through the Process API, not from the Document Inbox UI. It's useful when:

• Documents are already hosted online (your server, S3, Google Drive, etc.).
• You use Zapier, Make, or other automation tools that work with URLs.
• You have scripts or integrations that call the API.

How it works

1. You have document URLs that point directly to files (PDF, images, Office docs, etc.). The URLs must be publicly accessible without cookies or sign-in so PARSER can fetch them.

2. You call the Process API with a parser ID or classifier ID and your API key, then pass an array of URLs in the request body.

3. PARSER downloads each file. Parser URL imports are processed by that parser. Classifier URL imports are sorted first, then routed to the matching parser.

Processed documents show up in the same inbox as uploads and email. You can filter, edit, and export them as usual.

PARSER queues the batch only after every file has been stored. If one transfer fails before queueing, retry the complete request rather than assuming earlier files were accepted.

Where to use it

• API - Use the Process API (/api/v1/process/urls) with a urls array and exactly one of parserId or classifierId. Legacy parser-only URL imports still work at /api/v1/process/{parserId}/urls. See the API Reference for authentication and request format.
• Zapier - If your Zapier integration supports "Process from URL," you can send document URLs there instead of file uploads.
• Custom integrations - Any tool that can send HTTP requests can use the API to process documents from URLs.

Supported files and limits

The same file types as uploads are supported (PDF, PNG, JPEG, Word, Excel, etc.). The parser or classifier accepted file types setting still applies. Split behavior follows the parser or classifier split settings when the document is a PDF, including Split with AI.

Each request can contain up to 20 URLs, with a maximum of 100 MB per file. Each download must finish within five minutes.

Every URL and redirect must:

• Use HTTP or HTTPS
• Resolve to a public internet address
• Return the file without browser cookies, login credentials, or PARSER authorization
• Follow no more than five redirects

URLs for localhost, private networks, local devices, cloud metadata services, or files that require credentials are rejected.

If Document type rules are enabled, PARSER checks each URL import before extraction. Depending on the selected rule, PARSER can process only matching document types or skip matching types and process everything else. Skipped documents stay visible in the Document Inbox.

Tips

• Ensure URLs are publicly accessible and return the actual file (not a login page).
• Use HTTPS when possible. PARSER checks each redirect before downloading the file.
• If a batch fails during download, correct the failing URL and retry the complete batch.
• Use Document Type Rules if URL imports may include files that do not belong in this parser.
• Use Document Classifiers if URL imports may contain files for several parsers.

Next steps

• API Reference - Process API details and authentication
• Document Classifiers - Route mixed URL imports to different parsers
• Uploading Documents - Upload from the app
• Send Documents by Email - Send documents by email