Similarity search tool for Researcher

Want to have the Researcher agent find needed files even better? Provide it with a Semantic Search (RAG) tool!

This tool allows the agent to search files by text query (similar to how you find articles in Google), which is compared with file descriptions.

Set up

1. Create a vector database with descriptions of your project files by running:

python rag/write_descriptions.py

2. Get api key for Cohere

3. Add it to your .env file: COHERE_API_KEY=your_api_key_here

That’s all! Now your Researcher agent is equipped in additional tool and will be able to navigate big project better.