API Reference

Status: Template - to be completed when FastAPI routes are implemented (expected M4-M6)

---

Overview

The AI Orchestrator exposes a RESTful API for document processing, field extraction, and CRM integration. This page will document all available endpoints once the API is implemented.

---

Authentication

Parameter	Value
Method	TBD
Header	TBD
Token Format	TBD

---

Document Processing (WP2/WP3)

Upload Document

Get Extraction Results

Get Processing Status

---

Field Mapping (WP4)

Map to Schema

List Supported Schemas

Validate Mapping

---

Document Pre-filling (WP5)

Pre-fill Form

Supported Formats

---

System

Health Check

Model Info

---

Error Codes

Code	Description
400	Bad Request - Invalid input format
401	Unauthorized - Invalid or missing authentication
404	Not Found - Document or resource not found
413	Payload Too Large - Document exceeds size limit
422	Unprocessable Entity - Valid format but cannot process
429	Too Many Requests - Rate limit exceeded
500	Internal Server Error

---

Rate Limits

---

Known Architecture (from architecture.md)

The system architecture defines these processing stages that will map to API endpoints:

1. Document Digitization - OCR + text extraction
2. Field Extraction - LLM-based entity recognition
3. Schema Mapping - Zero-shot field mapping to target systems
4. Form Pre-filling - Automated population of output documents

See Architecture for the full system design.

---

This document will be populated as API endpoints are implemented during WP2-WP5 development.