Orchard
Every Apple needs an Orchard.
Local inference on Apple Silicon. The engine underneath Proxy. Run multiple models at once on your Mac.
Download Proxy
Read the Docs
Requires macOS 15+
Local First
- Multiple models, at the same time
- Structured output, multimodal, streaming
- Custom kernels and inference engine built for Apple Silicon
Open Source
- Private, offline, on device
- Python SDK on PyPI
- Rust SDK on crates.io