Origami

Local AI research assistant. Upload PDFs, take Markdown notes, and chat with an LLM that has context over your documents and notes.

Stack: FastAPI + LangGraph + ChromaDB + Ollama (backend) | Next.js 16 + React 19 + AI SDK v6 (frontend)

Prerequisites

• Ollama installed and running
• Bun (v1.3+)
• Python 3.13+
• uv (Python package manager)

Gallery

Light Theme:

Dark Theme:

Setup

1. Ollama

Install Ollama and pull the default model:

# macOS
brew install ollama

# Start the Ollama server (runs on port 11434)
ollama serve

# In a separate terminal, pull the default model
ollama pull deepseek-r1:8b

Make sure ollama serve is running before starting the backend.

2. Backend

cd backend

# Create virtual environment and install dependencies
uv sync

# Copy the example env file and adjust if needed
cp .env.example .env.local

# Start the FastAPI server (runs on port 8000)
uv run uvicorn main:app --reload

3. Frontend

cd frontend

# Install dependencies
bun install

# Copy the example env file and adjust if needed
cp .env.example .env.local

# Start the dev server (runs on port 3000)
bun run dev

4. Open the app

Visit http://localhost:3000.

Configuration

All config is driven by .env.local files (git-ignored). Each directory ships a .env.example that you copy and edit. Only edit .env.local — never commit it. If you add a new variable, update .env.example too so others can see what's available.

Backend (backend/.env.local)

Variable	Default	Description
OLLAMA_URL	http://localhost:11434	Ollama API endpoint
OLLAMA_MODEL	deepseek-r1:8b	LLM used for chat, agent, and ingestion
OLLAMA_TIMEOUT	120	Request timeout in seconds
EMBEDDING_MODEL	bge-small-en-v1.5	Sentence-transformer for ChromaDB (re-ingest if changed)
CHROMA_DIR	chroma_data	Path to ChromaDB storage
CHROMA_COLLECTION	documents	ChromaDB collection name
FRONTEND_URL	http://localhost:3000	Allowed CORS origin
CHUNK_SIZE	1200	Characters per chunk during ingestion
CHUNK_OVERLAP	300	Overlap between chunks

Frontend (frontend/.env.local)

Variable	Default	Description
NEXT_PUBLIC_API_URL	http://localhost:8000	Backend API URL

Switching models

ollama pull qwen2.5:7b
# then in backend/.env.local:
OLLAMA_MODEL=qwen2.5:7b
# restart the backend

Adding a new env variable

1. Add it to backend/config.py with an os.getenv() default
2. Add it to backend/.env.example with a comment
3. Add it to your local backend/.env.local

Project Structure

Origami/
├── backend/              # FastAPI + Python
│   ├── main.py           # App entry, CORS, routers
│   ├── config.py         # Centralized env var config
│   ├── .env.example      # Template for .env.local
│   ├── routes/           # API endpoints (chat, upload, notes, documents)
│   ├── services/         # Ollama, ChromaDB, RAG, ingestion
│   ├── prompts/          # LLM prompt templates
│   ├── notes/            # Markdown note files
│   └── pyproject.toml
├── frontend/             # Next.js + React
│   ├── app/              # Pages and API routes
│   ├── components/       # Chat, sidebar, UI components
│   ├── lib/api/          # Backend API client
│   ├── .env.example      # Template for .env.local
│   └── package.json
└── package.json          # Root workspace