🧠 AI Document Analysis API

An intelligent document processing API that extracts, analyses, and summarises content from PDF, DOCX, and image files using Claude AI.

Features

Feature	Detail
Multi-format support	PDF, DOCX, JPEG, PNG, TIFF, BMP, WEBP, GIF
Text extraction	PyMuPDF (PDF), python-docx (DOCX), Tesseract OCR (images + scanned PDFs)
AI Summarisation	Concise 2–5 sentence summaries via Claude claude-opus-4-5
Entity extraction	PERSON, ORGANIZATION, LOCATION, DATE, MONEY, PRODUCT, EVENT
Sentiment analysis	positive / negative / neutral with confidence score
Authentication	API key via X-API-Key header

---

API Reference

POST /analyse

Upload a document for processing.

Headers:

X-API-Key: <your-api-key>
Content-Type: multipart/form-data

Body:

file: <file upload>

Response:

{
  "filename": "report.pdf",
  "file_type": "PDF",
  "num_pages": 5,
  "word_count": 1842,
  "extracted_text": "...",
  "summary": "This report analyses...",
  "entities": [
    {"text": "John Smith", "type": "PERSON"},
    {"text": "Acme Corp", "type": "ORGANIZATION"},
    {"text": "$2.5 million", "type": "MONEY"}
  ],
  "sentiment": {
    "label": "positive",
    "score": 0.87
  },
  "metadata": {"num_pages": 5}
}

GET /health

Returns {"status": "ok"}.

GET /

Returns API info and endpoint list.

---

Quick Start (Local)

# 1. Clone
git clone https://github.com/YOUR_USERNAME/docai.git
cd docai

# 2. Install dependencies
pip install -r requirements.txt

# 3. Set environment variables
export GROQ_API_KEY=sk-ant-...
export DOC_API_KEY=your-secret-key

# 4. Run
uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload

Docker

docker build -t docai .
docker run -p 8000:8000 \
  -e GROQ_API_KEY=sk-ant-... \
  -e DOC_API_KEY=your-secret-key \
  docai

Deploy to Railway (Free Tier)

1. Push this repo to GitHub
2. Go to railway.app → New Project → Deploy from GitHub
3. Select your repo
4. Add environment variables:
   • GROQ_API_KEY = your GROQ key
   • DOC_API_KEY = your chosen API key
5. Deploy — Railway will auto-detect the Dockerfile

Deploy to Render (Free Tier)

1. Push repo to GitHub
2. Go to render.com → New → Web Service
3. Connect your GitHub repo
4. Set:
   • Runtime: Docker
   • Environment Variables: GROQ_API_KEY, DOC_API_KEY
5. Deploy

---

Test with curl

# PDF
curl -X POST https://YOUR-DOMAIN/analyse \
  -H "X-API-Key: your-secret-key" \
  -F "file=@document.pdf"

# DOCX
curl -X POST https://YOUR-DOMAIN/analyse \
  -H "X-API-Key: your-secret-key" \
  -F "file=@report.docx"

# Image
curl -X POST https://YOUR-DOMAIN/analyse \
  -H "X-API-Key: your-secret-key" \
  -F "file=@scan.jpg"

---

Environment Variables

Variable	Description	Default
GROQ_API_KEY	GROQ API key (required)	—
DOC_API_KEY	Your chosen API key for auth	docai-secret-key-2024
PORT	Server port (auto-set by Railway/Render)	8000

---

Architecture

Client ──► POST /analyse (multipart file)
              │
              ▼
         FastAPI Router
              │
              ▼
         DocumentProcessor
         ├── PDF  → PyMuPDF → text extraction
         │          └── scanned? → Claude Vision OCR
         ├── DOCX → python-docx → paragraph/table text
         └── Image → Tesseract OCR
              │
              ▼
         Claude claude-opus-4-5 (single prompt)
         ├── Summary
         ├── Entities (NER)
         └── Sentiment
              │
              ▼
         JSON Response

Tech Stack

• FastAPI — async Python web framework
• Claude claude-opus-4-5 — AI summarisation, NER, sentiment, OCR
• PyMuPDF — PDF text extraction
• Tesseract + pytesseract — Image OCR
• python-docx — DOCX text extraction
• Docker — containerised deployment

AI used

Claude and Perplexity