Turn documents into APIs
Extract a PDF and get a clean REST endpoint anyone can POST questions to. No API keys, no auth headers, no tokens in query strings.How it works
Extract your document
Upload via web app, CLI, or SDK. OkraPDF extracts all text, tables, and figures into a queryable store.
Create a public link
Generate a share link with
role: 'ask'. OkraPDF creates a human-readable slug from your filename.Clean, shareable URLs
Every ask link gets a human-readable slug derived from the document filename:| Document | Public endpoint |
|---|---|
amcor-2023q4-earnings.pdf | .../v1/documents/amcor-2023q4-earning-a3f8b2/completion |
NVIDIA 10-K Report.pdf | .../v1/documents/nvidia-10-k-report-c7d1e4/completion |
invoice-march.pdf | .../v1/documents/invoice-march-8b3f2a/completion |
Access controls
You stay in control of every endpoint you create:Usage limits
Set
maxViews to cap total requests. Link expires automatically when the limit is reached.Kill switch
Revoke any link instantly. Returns
410 Gone on the next request.TTL
Set
expiresInMs for time-limited access. Demo for a client call, then it’s gone.Rate limiting
10 requests per minute per slug, built in. No configuration needed.
Analytics
Track usage per link:Use cases
Embed Q&A in your product
Embed Q&A in your product
Your backend creates the link once. Your frontend calls it directly — CORS is set to
*.Demo endpoint for sales calls
Demo endpoint for sales calls
Create a link with
maxViews: 50 and expiresInMs: 86400000 (24h). Share the curl command in the meeting. It stops working after the call.Client-facing document chatbot
Client-facing document chatbot
Upload a client’s report, create an ask link, share the URL. They can ask questions about their own document without needing an OkraPDF account.
Internal knowledge base
Internal knowledge base
Upload policy documents, create ask links, embed them in Slack bots or internal tools. Employees query documents with plain HTTP.