askme - Codebase RAG CLI

A Command Line Interface (CLI) tool for local Retrieval-Augmented Generation (RAG) over your codebase. It scans, indexes, and lets you chat with your project using local LLMs (via an OpenAI-compatible API) and a Qdrant vector database, returning context-aware answers with source citations.

The Python package is named askme and exposes a single askme console script.

Quickstart

Prerequisites

• Python 3.11+
• An OpenAI-compatible LLM and embeddings endpoint (e.g. Ollama)
• Qdrant vector database — either:
  • server mode: running Qdrant instance (Docker/remote)
  • local mode: file-based store on disk, no server required (quick start)

1. Prepare models (example with Ollama)

ollama pull llama3
ollama pull embeddinggemma:300m

2. Install

# Using uv (recommended)
uv sync
uv run askme

# Or with pip
pip install .
askme

3. Run

Inside the codebase you want to query:

askme

On first launch, the tool guides you through configuration, scans the project, and builds the Qdrant collection. Subsequent runs reuse the existing index and only re-embed changed files.

Key Commands

Configuration

On first run inside a project directory, askme prompts for configuration and stores it in ./.cfg/settings.json. Conversation histories are stored under ./.cfg/history/.

Key options:

• qdrant_mode: server (remote/Docker) or local (file-based)
• qdrant_local_path: on-disk path for local mode (default ./data/vector_store)
• llm_api_base: OpenAI-compatible LLM API base URL
• embedding_model: embedding model name
• chunk_size / chunk_overlap: chunking parameters
• file_extensions / excluded: file and directory filtering

Documentation

Full documentation with setup details, storage schema, development guide, and contributing guidelines is available in /doc directory.