Own your AI conversations. Scry is a self-hosted conversation archive that keeps your AI chat history private, searchable, and portable across providers.

Your conversations contain personal thoughts, business ideas, and sensitive information that have no business being hosted by "big tech" on the public internet. Scry keeps everything local on your own machine—no cloud services, no telemetry, no data leaving your infrastructure. Your PostgreSQL database, your data, your control.

Don't let your conversation history be locked into one AI provider. Scry imports from ChatGPT, Claude, and OpenWebUI, giving you a unified archive that survives provider switches, account changes, or service shutdowns.

Seamlessly sync conversations with Open WebUI—pull your history into Scry for archival, or export conversations back to OpenWebUI. Use OpenWebUI as your AI frontend while Scry serves as your permanent, searchable archive.

• Multi-Source Import: ChatGPT JSON exports, Claude conversations, OpenWebUI API sync
• Smart Search: Combined full-text and semantic vector search for finding exactly what you need
• RAG Integration: Contextual retrieval for AI-powered search with surrounding conversation context
• Background Sync: Automatic OpenWebUI synchronization with change detection
• Self-Contained: Runs entirely on your own hardware with Docker

# Pull the image
docker pull ghcr.io/mbrichman/Scry:main

# Configure environment
cp .env.example .env
# Edit .env with your PostgreSQL credentials

# Start the stack
docker compose up -d

# Access the application
open http://localhost:5001

Create a .env file (use .env.example as a template):

# PostgreSQL credentials
POSTGRES_USER=Scry
POSTGRES_PASSWORD=your-secure-password
POSTGRES_DB=Scry

# Application settings
SECRET_KEY=your-secret-key-here

To sync conversations from OpenWebUI:

1. Go to Settings in Scry
2. Enter your OpenWebUI URL (e.g., https://your-openwebui-instance)
3. Add your OpenWebUI API key
4. Click Sync to pull conversations

Scry tracks sync state and only fetches new or updated conversations on subsequent syncs.

Scry/
├── api/                  # REST API endpoints
├── controllers/          # Request handlers
├── db/
│   ├── models/          # SQLAlchemy models
│   ├── repositories/    # Data access layer
│   ├── services/        # Business logic
│   │   ├── search_service.py          # Hybrid FTS + vector search
│   │   ├── import_service.py          # Multi-format import
│   │   ├── sync_service.py            # OpenWebUI sync
│   │   └── contextual_retrieval_service.py  # RAG context windows
│   └── workers/         # Background embedding generation
├── templates/           # Web UI (Jinja2)
└── tests/               # Test suite

• Backend: Flask, SQLAlchemy
• Database: PostgreSQL 17 with pgvector
• Search: PostgreSQL FTS + sentence-transformers embeddings
• Deployment: Docker & Docker Compose

# Run tests
pytest

# Run specific test categories
pytest tests/unit/
pytest tests/integration/

# With coverage
pytest --cov=. --cov-report=html

1. Create a feature branch from main
2. Make your changes
3. Ensure tests pass: pytest
4. Push and create a pull request

MIT License - see LICENSE for details.