Alternatives to Pmetal

Run open AI models locally with one command, with per-token Ollama Cloud models when local hardware runs out of headroom.

Run open LLMs locally with LM Studio, then let the Bionic agent loose on documents, code and automations.

LFM2.5 open-weight models run on-device LLM inference with text, vision, and speech on CPUs, NPUs, and Apple Silicon.

Microsoft's open-source C++ inference framework that runs 1-bit and ternary BitNet b1.58 models locally on CPUs, plus an official GPU kernel.

Free, MIT-licensed local LLM inference for Apple Silicon Macs, with SSD-backed prefix caching and native OpenAI and Anthropic APIs.

On-device LLM inference for Apple Silicon: uzu runtime, architecture-co-designed quantization, and models tuned for sub-200ms interactivity.

Free open-source local AI server that runs LLMs, image, music, video, and 3D generation on your own Apple Silicon Mac.

AgenticSeek is a fully local, open-source AI assistant that browses the web, writes and runs code, and plans multi-step tasks on your own machine.

Open-source local AI workspace to run and train text, image, video and audio models on your own hardware

Free OpenClaw Book: a 45-chapter GitBook on how the OpenClaw AI assistant works, down to the source line

Enclave runs open-source AI models locally on iPhone and Mac for private offline AI chat, voice, and document Q&A.

A headless React hooks library that runs Vicuna-13B chat entirely in your browser via WebGPU, with no prompt or completion leaving the client.

Free, open-source desktop app for running and fine-tuning RWKV RNN language models locally with effectively unbounded context.

Run LLMs, speech-to-text, and image generation fully offline on your device via a Kotlin Multiplatform library, an app, and an IDE plugin.

Run local HuggingFace models or route prompts to Claude, GPT and Gemini with your own keys in this free, open-source AI desktop client

Run private local AI agents on your Mac with Osaurus — free, MIT-licensed, cloud-optional.

Free .aimodel bundles plus rerunnable conversion recipes for Apple's Core AI runtime on iOS 27 and macOS 27.

Run a local AI agent on your own GPU: llama.cpp model, GBNF tool calls, zero token bill.

picoLLM runs quantized LLMs fully on-device — local inference for phones, embedded Linux, and microcontrollers with no cloud calls.

LLM Hub runs 15+ AI models fully offline on Android and iOS for private on-device chat, image, video and code generation.

Atomic Chat runs 1,000+ open models fully offline on your Mac, PC, Linux box or phone, with no account and no message caps.

Benchmark local LLM tokens-per-second through Ollama with a single CLI command and a public results table.

OnnxStream streams ONNX model weights from disk so large models like SDXL and Mistral 7B run on tiny-RAM devices.

Open-source, local-first terminal assistant that turns plain English into shell commands and runs fully offline through Ollama.

AI research lab building tiny sub-1 MB models that run on microcontrollers and edge devices.

Free Podman Desktop extension for running open source LLMs locally, with recipes, a model catalog and OpenAI-compatible endpoints.

Run any GGUF AI model locally on Android with zero network permissions

Iris Android runs GGUF language models offline on your phone using llama.cpp — no server, no account, no data leaving the handset.

Run open-source LLMs fully offline on Mac and Windows with an encrypted local vault

QVAC SDK: one open-source API for on-device AI across text, vision and audio, with no cloud