LocalMind

LocalMind is an local LLM Chat App fully compatible with the Model Context Protocol. It uses Azure OpenAI as a LLM backend and you can connect it to all MCP Servers out there.

Local Development

Create a .env file in the backend folder:

APP_CONFIG_FILE_PATH=config.yaml
AZURE_OPENAI_API_KEY=x
AZURE_OPENAI_DEPLOYMENT=x
AZURE_OPENAI_ENDPOINT=https://x.openai.azure.com
AZURE_OPENAI_API_VERSION=2024-07-01-preview
AZURE_OPENAI_CHAT_MODEL=gpt-4o
AZURE_OPENAI_EMBEDDINGS_MODEL=embedding

Create a config.yaml file in your backend folder:

server:
- name: [SERVER_NAME]
  command: [SERVER_COMMAND]
  args:
  - [SERVER_ARGS]
[...]

To work on the frontend in browser with the python backend up and running:

./dev.sh frontend-dev

To run the Tauri App in development mode with the python backend:

./dev.sh app-dev

RAG MCP Server

If you would like to use or work on the RAG MCP Server, first create a .env file in the rag folder:

AZURE_OPENAI_API_KEY=x
AZURE_OPENAI_DEPLOYMENT=x
AZURE_OPENAI_ENDPOINT=https://x.openai.azure.com
AZURE_OPENAI_API_VERSION=2024-07-01-preview
AZURE_OPENAI_CHAT_MODEL=gpt-4o
AZURE_OPENAI_EMBEDDINGS_MODEL=embedding

Create venv and install dependecies:

cd rag
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Then add the following config entry to your config.yaml in your backend folder:

server:
- name: rag
  command: [ABSOLUTE_PATH]/rag/.venv/bin/python3
  args:
  - [ABSOLUTE_PATH]/rag/main.py

Important

Currently only works with Azure OpenAI Service.