| title | Snapr |
|---|---|
| emoji | 🏷️ |
| colorFrom | yellow |
| colorTo | blue |
| sdk | docker |
| app_port | 8080 |
| pinned | false |
| short_description | Multi-agent AI that finds the best online deals. |
📷 Screenshots
Snapr is an AI-powered system that finds the best online deals using a pipeline of collaborative agents powered by LLMs, traditional ML models, embeddings, and retrieval-augmented generation (RAG).
- Central agent orchestrates deal processing with specialized agents
- Real-time deal fetching and price prediction using LLMs and traditional models
- Automatic filtering and memory to avoid duplicates
- Remote execution on Modal with live activity logs
- DevOps: Docker, CI/CD with GitHub Actions, pre-commit hooks
- Unit testing, security checks, and code quality enforcement
- MkDocs-based documentation
👉 For full details, see Core Features and Development & DevOps..
To set up locally and install the app, see the "Local Deployment" section in the technical documentation.
- Single Container Deployment: FastAPI serves both the Gradio app and documentation
- Main app at
/- Interactive Gradio interface - Documentation at
/docs- MkDocs static site - No reverse proxy needed - everything runs through uvicorn/FastAPI
- Main app at
- Containerized with Docker for consistent deployment across environments
-
Documentation includes:
- Technical and functional docs
- Schematics of technical architecture, functional workflow, and user flow
MIT
