Back to Home
matthewhanson icon

athenaeum

Verified Safe

by matthewhanson

Overview

A RAG system providing vector search and LLM-powered chat/question answering over markdown documents via a REST API, with optimized serverless deployment to AWS Lambda.

Installation

Run Command
uv run athenaeum serve --index ./index

Environment Variables

  • OPENAI_API_KEY
  • CDK_DEFAULT_ACCOUNT
  • CDK_DEFAULT_REGION
  • ATHENAEUM_INDEX_DIR
  • CHAT_SYSTEM_PROMPT
  • CHAT_SYSTEM_PROMPT_FILE
  • OAUTH_JWKS_URL
  • OAUTH_ISSUER
  • OAUTH_AUDIENCE

Security Notes

The server uses `allow_origins=["*"]` for CORS by default, which is explicitly noted as needing restriction in production environments. All sensitive keys like `OPENAI_API_KEY` and OAuth configuration are designed to be loaded from environment variables, avoiding hardcoded secrets. The use of `subprocess.Popen` and binding to `0.0.0.0` in the CLI `serve` command is intentional for development/container usage and does not expose immediate vulnerabilities given its context. The `oauth_authorizer.py` provides robust JWT/OAuth token validation.

Stats

Interest Score0
Security Score8
Cost ClassMedium
Avg Tokens1500
Stars0
Forks0
Last Update2025-12-15

Tags

RAGLLMMarkdownFastAPILlamaIndexAWS LambdaVector Search