InferrLM
A mobile app for Android and iOS that runs language models directly on the device and turns the phone into a local inference server, with optional cloud model support.
InferrLM is a mobile application that runs large and small language models directly on an Android or iOS device, and can expose the device as a local inference server so other devices on the same network reach it over REST APIs. Core inference runs through llama.cpp with GGUF models on both platforms, plus MLX inference on Apple Silicon and Apple's own Foundation Model on supported devices; cloud models from OpenAI, Gemini, and Anthropic are also supported when the user provides their own API keys and an InferrLM account.
It supports vision and multimodal input through models with a matching projector file, and a built-in camera can capture images to send directly to a model. Document processing includes on-device OCR of attached files and retrieval-augmented generation over ingested documents, and customizable base URLs let it reach OpenAI-compatible providers such as OpenRouter, Groq, Ollama, LM Studio, or Together AI.
It is documented in eight languages and was previously named Inferra.