UnieAI Parser Overview
Parse PDFs and images through stateless APIs, then use separate knowledge-base ingestion and Runtime APIs for persistence, search, and RAG.
UnieAI Parser provides synchronous, stateless APIs for parsing a PDF or image. Knowledge-base ingestion reuses parsing capabilities but is a separate background workflow that persists files, builds chunks and indexes, and makes the content available to search and RAG.
In this tab
Quickstart
Authenticate and run your first parse, search, and chat call over HTTP.
Document Parsing API
Stateless PDF / image parsing endpoints and response format.
Chunking Strategies
Semantic chunking, table atomicity, and figure recovery.
Client options
Use any HTTP client, such as server-side fetch, Python requests, or curl. Do not embed long-lived API keys in browser code. OpenAI-compatible clients apply only to compatible /v1 endpoints such as Chat Completions and Embeddings, not to Parser or knowledge-base search endpoints.
Endpoints at a glance
| Capability | Endpoint |
|---|---|
| Parse PDF | POST /api/parse/pdf |
| Parse image | POST /api/parse/image |
| KB search | POST /v1/kb/search |
| Chat (OpenAI-compatible) | POST /v1/chat/completions |
| Embeddings | POST /v1/embeddings |
Continue with the Quickstart to start integrating.