Hacktoberfest 2026 is here, and BOOTH is open to contributors!

I’ve been building BOOTH, a lightweight checkpoint layer for LLM outputs.

The idea is simple: don’t automatically trust every LLM response. Check it first.

With check_with_evidence(), BOOTH lets you compare an LLM’s answer against evidence your RAG pipeline has already retrieved.

BOOTH v0.5.2

• Zero runtime dependencies
• Provider-agnostic
• Sync + async APIs
• Structured results
• 300+ tests
• MIT licensed

I’m looking for developers interested in LLM reliability, RAG, and Python to contribute this Hacktoberfest.

🔗 Explore BOOTH: https://github.com/Vedantgitbot/booth
🛠️ Find an issue: https://github.com/Vedantgitbot/booth/issues

Whether it’s your first open-source contribution or your hundredth, you’re welcome!