papertrail

Track ArXiv papers by topic and get AI-generated summaries. A FastAPI backend that fetches papers, generates concise summaries using Claude, and serves everything through a REST API.

Features

• ArXiv integration: Fetch papers by search query with full metadata parsing
• AI summaries: Generate 2-3 sentence technical summaries via Claude
• Topic tracking: Save search queries and run automated digests
• REST API: Clean FastAPI endpoints with auto-generated OpenAPI docs
• SQLite storage: Lightweight persistence with upsert logic
• Tested: Unit tests for feed parsing and data models

Quick Start

# Clone
git clone https://github.com/zgroves24/papertrail.git
cd papertrail

# Set up environment
python -m venv .venv
source .venv/bin/activate  # or .venv\Scripts\activate on Windows
pip install -e ".[dev]"

# Configure
cp .env.example .env
# Add your Anthropic API key to .env

# Run
python -m src.main

API docs at localhost:8000/docs.

API Endpoints

Method	Path	Description
GET	/papers	List tracked papers (paginated)
POST	/papers/fetch	Fetch papers from ArXiv by query
POST	/papers/summarize	Generate AI summaries for unsummarized papers
GET	/topics	List tracked topics
POST	/topics	Add a new topic to track
POST	/digest	Run full digest: fetch + summarize for all topics

Example Usage

# Fetch recent transformer papers
curl -X POST "localhost:8000/papers/fetch?query=transformer+attention&max_results=5"

# Add a topic to track
curl -X POST localhost:8000/topics \
  -H 'Content-Type: application/json' \
  -d '{"name": "LLM Agents", "query": "large language model agents"}'

# Run a full digest
curl -X POST localhost:8000/digest

Testing

pytest

Tech Stack

• FastAPI for the REST API
• Anthropic SDK for AI summaries
• httpx for async HTTP
• SQLite for storage
• Pydantic for data validation
• ruff for linting

License

MIT