LiteRT-LM
Google
Models, data & tools
Google’s production-oriented orchestration and inference layer for running LLMs and multimodal models with LiteRT across Android, iOS, web, desktop and IoT. It supports acceleration, function calling and models including Gemma, Llama, Phi and Qwen.
Free · Open source
Self-hosted / on-premise · Edge / local · Downloadable
Edge deployment runtimes
LiteRT-LM
build-integrate-ai
agrifood-data-and-ai-infrastructure
buildable-or-downloadable-component
edge-deployment-runtimes
Free · Open source
Self-hosted / on-premise · Edge / local · Downloadable
2026-08-26
- Run an LLM offline on a field phone
- Deploy a VLM on Raspberry Pi
- Add local function calling to an agrifood app
- Use GPU or NPU acceleration at the edge
- Package one model across Android iOS and desktop
- Keep farm data on the device