Recall

Local notes search with a small RAG stack. Recall indexes your markdown/text notes and lets you query them using a local Ollama model (no cloud, no API keys).

Setup

1. Install Ollama from ollama.com and make sure it’s running (ollama serve if needed).

2. From the project root:

   ./setup.sh
   ./run.sh

3. Open http://localhost:8000 in your browser.

Usage

Commands

Command	Description
/help	Show available commands
/config	Display current configuration
/sources	List all indexed files
/clear	Clear conversation history

Example Queries

what notes do I have about Python?
summarize my learning notes
what projects have I been working on?

---

Project Structure

src/
├── server.py          # FastAPI backend (API + static serving)
├── app.py             # RAG chain setup
├── config.py          # Configuration (edit note paths here)
├── data_ingest.py     # Load notes from directories
├── chunker.py         # Text splitting
├── vector_store.py    # FAISS operations
├── static/
│   ├── index.html     # Terminal-style UI
│   ├── style.css      # Terminal theme
│   └── script.js      # Frontend logic
└── data/processed/    # Auto-generated vector index

Configuration

All settings in src/config.py:

Setting	Default	Description
NOTES_PATHS	~/Documents	Directories to index
NOTES_EXTENSIONS	.md, .txt	File types to include
MODELS['llm']	llama3.2:3b	Ollama model name
MODELS['top_k']	5	Results per query
MODELS['chunk_size']	500	Chunk size in chars
OLLAMA_BASE_URL	localhost:11434	Ollama server address

Tech Stack

Component	Technology
LLM	Ollama (llama3.2:3b, local)
Embeddings	all-MiniLM-L6-v2 (local)
Vector Store	FAISS
Framework	LangChain
UI	FastAPI + HTML/CSS/JS

License

MIT