[ ABORT TO HUD ]
SEQ. 1
SEQ. 2
SEQ. 3
SEQ. 4
SEQ. 5
Foundry Local v1.1–1.2
Run AI Models Locally
Foundry Local enables developers to run AI models directly on their own hardware for offline scenarios, edge computing, and low-latency applications. The v1.1 and v1.2 releases (early 2026) significantly expanded platform and model support.
What's New in Foundry Local v1.1–1.2
| Feature | Details |
|---|---|
| Linux ARM64 Support | Run Foundry Local on ARM64-based Linux devices (Raspberry Pi, NVIDIA Jetson, etc.) |
| Live Audio Transcription | Real-time speech-to-text processing directly on-device |
| Text Embeddings | Generate vector embeddings locally for offline RAG pipelines |
| Qwen 3.5 Vision Support | Run Qwen 3.5 Vision model locally for on-device multimodal inference |
| ONNX Runtime 1.26 | Latest ONNX Runtime for optimized model execution across hardware |
Supported Languages
Foundry Local provides SDKs for Python, JavaScript, C#, and Rust, making it accessible across a wide range of development ecosystems.
💡 Key Insight: Foundry Local is ideal for scenarios requiring data sovereignty, air-gapped environments, or ultra-low latency. Use it alongside cloud-based Foundry for a hybrid AI architecture.
⌨ HANDS-ON LABRun a Model On-Device with Foundry Local
⭐ +150 XPNo cloud, no keys - run a model on your own silicon. Install Foundry Local, then chat with Phi entirely offline.
1Install Foundry Local using winget.
2Run a local model with the foundry CLI (e.g. phi-4-mini). Hardware-optimal variant auto-selected.
OBJECTIVE 1 / 2 — type "hint" if stuck
FOUNDRY VERIFICATION
QUERY 1 // 1
What new platform support did Foundry Local v1.1–1.2 add?
Windows ARM only
Linux ARM64 support
macOS Intel only
Android