customizable-chatbot

company-chatbot implementation using RAG(Retrieval Augmented Generation)

Bot able to answer question related to information present in the Doc ->

Tip: chromadb@1.7.3 works well with LangChain’s camelCase request body.

Run a local Chroma server (Python, isolated venv)

python3 -m venv .venv && source .venv/bin/activate pip install 'chromadb==0.4.24' 'numpy<2' /Users/you/path/.venv/bin/chroma run
--host 0.0.0.0 --port 8001
--path /absolute/path/to/chroma_db

check if it is running?

curl -s http://127.0.0.1:8001/api/v1/heartbeat

Why these versions:

Server: 0.4.24 exposes v1 REST routes used by LangChain’s Chroma integration. NumPy < 2 fixes the np.float_ import error for 0.4.x. JS client 1.7.3 honors LangChain’s camelCase request fields.

What is .venv here for ?

It’s a Python virtual environment for running the Chroma server.

It isolates Python packages (e.g., chromadb 0.4.24, numpy<2) from your system/global Python so versions don’t conflict with your Node app or other projects.

You activate it to start/stop the server:

Activate: source /Users/jigyasaupadhyay/Documents/Work/customizable-chatbot/.venv/bin/activate

Run Chroma: /Users/jigyasaupadhyay/Documents/Work/customizable-chatbot/.venv/bin/chroma run ...

Deactivate: deactivate