Ragit – chat with any folder of documents using a local LLM

Local RAG CLI to chat with any folder of documents using Ollama.

cd ~/ragit
python3 -m pip install -e .

If your default Python is 3.14+, use Python 3.10–3.13 (recommended: 3.12) because some vector DB dependencies may not publish wheels for very new Python versions yet.

Make sure Ollama is installed and running:

ollama pull nomic-embed-text
ollama serve

Index a folder:

ragit index ./docs

Start chat:

ragit chat ./docs

List available Ollama models:

ragit models

Clear an index:

ragit clear ./docs

ragit

implements Retrieval-Augmented Generation (RAG):

• It loads supported documents ( .txt

,.md

,.pdf

,.docx

) recursively. - It splits text into overlapping chunks (about 500 words with 50-word overlap).

• It creates embeddings using Ollama ( nomic-embed-text

) and stores vectors in local ChromaDB at~/.ragit/<hash_of_path>/

. - During chat, it embeds each query, retrieves the top relevant chunks, and injects them into a prompt.

• It streams an answer from a local Ollama chat model (prefers llama3.2

if available), then shows source chunks used.

• All data stays local on your machine (Ollama + Chroma local persistence).
• Indexes are stored under ~/.ragit/<hash_of_path>/

. - Files that cannot be parsed are skipped with a clear error message.