*What it does:* - Semantic search across 617 FAR Part 52 clauses - Pre-vectorized with 384-dim embeddings (all-MiniLM-L6-v2) - Returns relevant clauses with similarity scores - Daily auto-updates from acquisition.gov - OpenAPI spec for AI agent integration
*Why it exists:* If you're building AI for government contracting, your LLM will hallucinate legal citations. A wrong FAR clause = disqualification. This solves that.
*Try it free:* https://blueskylineassets.github.io/far-rag-api/honeypot/
*API access (RapidAPI):* https://rapidapi.com/yschang/api/far-rag-federal-acquisition...
*Code (open source):* https://github.com/blueskylineassets/far-rag-api
Built with FastAPI + sentence-transformers. All data is public domain (17 U.S.C. § 105).
Open to feedback!