Alternatives to Pmetal

Run open models locally or in the cloud with one command — then plug them into Claude Code, Codex, and other coding agents.

Open-source, local-first AI client for private, multi-model conversations.

Deep technical guide to building a local-first AI assistant with OpenClaw

100% local, open-source AI assistant — private, free, and offline.

Microsoft's open-source framework for running 1-bit LLMs with fast, lossless CPU/GPU inference

Unified AI inference platform from kernel to cloud for any hardware, now with Qualcomm support and open-source Mojo.

SambaNova Cloud: fastest open-model AI inference with RDU hardware and agentic optimizations.

Open-source desktop AI agent with autonomous learning loop and deep memory

Open-weight on-device AI with native audio, vision, and Japanese models, free under $10M revenue.

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

Fastest offline AI server for Apple Silicon: local LLMs, images, music, video, 3D.

Self-improving inference API that routes every call to the best model and retrains itself from your traffic.

Run LLMs in-browser with WebGPU — headless React hooks, just useLLM().

Flat-rate, hyper-fast inference on open LLMs for agentic coding and production workloads.

Private on-device LLM, speech-to-text, and image generation for Kotlin Multiplatform.

Free, MIT-licensed desktop app for running text and image prompts across 11 AI providers side by side.

MAX Engine serves open-source LLMs through an OpenAI-compatible API on NVIDIA, AMD, and Apple silicon with no CUDA or PyTorch dependency.

Fine-tune and run LLMs locally with Unsloth — custom CUDA kernels cut VRAM and speed up training on your own GPU.

Free, private AI commit message generator for Apple Silicon Macs.

CoreWeave is an AI-native GPU cloud for large-scale model training, reinforcement learning, and low-latency inference.

Enterprise-managed LLM fine-tuning and serving platform by Rubrik

Managed Ray platform for distributed training, batch inference, and data curation at scale.

Run 123+ open-source models locally or connect online APIs in one free, open-source desktop app

DeepInfra is a serverless inference cloud serving 100+ open models — DeepSeek-V4-Flash-0731 at $0.06 per 1M input tokens — through one OpenAI-compatible API

Free local RAG chatbot for RTX GPUs – private document Q&A on your PC

Run private local AI agents with LM Studio Bionic

High-performance open-source LLM and multimodal inference serving.

High-throughput, memory-efficient open-source LLM inference and serving engine

Open-source desktop app for running RWKV RNN LLMs locally with infinite context.

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