SEC Insights is a full-stack Retrieval Augmented Generation (RAG) application that leverages LlamaIndex to answer questions about SEC filings (10-K, 10-Q). It's designed for developers looking to build production-ready RAG systems, offering a reference implementation with features like source citation, PDF viewing, API tool integration, and streaming responses.

How It Works

The application employs a RAG architecture, using LlamaIndex for document indexing and retrieval. It integrates with OpenAI's GPT models for generation and text-embedding-ada-002 for embeddings. A PGVector database stores the indexed data. The backend is built with FastAPI, and the frontend uses React/Next.js with Tailwind CSS. It also incorporates external tools like Polygon.io for quantitative data.

Quick Start & Requirements

• Installation: Recommended setup via GitHub Codespaces (.devcontainer/devcontainer.json) for pre-configured dependencies. Local setup requires Docker Compose.
• Prerequisites: Docker, OpenAI API key, potentially other API keys (e.g., Polygon.io).
• Resources: Local setup involves Docker containers. GitHub Codespaces offers a faster, pre-configured environment.
• Links: secinsights.ai, YouTube Tutorial

Highlighted Details

• Full-stack RAG example with production-ready features.
• Supports chat-based Q&A, source citation, PDF viewing, and API tool integration.
• Features token-level and reasoning step streaming via Server-Sent Events.
• Infrastructure-as-code for Vercel/Render deployment with CI/CD.

Maintenance & Community

• Active development with open contributions encouraged.
• Discord channel available for support: #sec-insights

Licensing & Compatibility

• License not explicitly stated in the README.

Limitations & Caveats

The frontend is not currently mobile-optimized. While functional, there is stated room for improvement in RAG performance.