YS0meone/Corvus

Multi-agent AI research system — finds academic papers via semantic search & citation snowballing, then answers questions over them using agentic RAG with self-reflection. Built with LangGraph, FastAPI, Celery, and Qdrant.

47
/ 100
Emerging

Employs a supervisor-subagent architecture where a top-level coordinator classifies queries, generates multi-step plans, and routes to specialized agents—Paper Finding (with iterative search, citation snowballing, and Cohere reranking) and Q&A (with scoped vector retrieval and self-evaluation loops). Integrates Semantic Scholar (200M+ papers), Tavily web search, Grobid PDF extraction, and OpenAI embeddings into a unified stack with Postgres checkpoints, Redis task queuing, and a React frontend; also exposes search capabilities via MCP server for Claude Desktop integration.

No Package No Dependents
Maintenance 10 / 25
Adoption 9 / 25
Maturity 15 / 25
Community 13 / 25

How are scores calculated?

Stars

84

Forks

10

Language

Python

License

MIT

Last pushed

Feb 28, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/rag/YS0meone/Corvus"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.