Magnitude
SHA-256An open source inference server that profiles your hardware, picks the best local models, and plugs into agents like Pi, OpenCode, Claude Code, and more.
Run your existing AI agents on local models automatically profiled and tuned for your hardware—free, private, and offline.
Core Features
- Hardware profiling and model recommendations with estimated tokens/sec
- One-prompt onboarding: your agent can set itself up
- On-demand model loading and automatic unloading when idle
- Fully private and offline: prompts, files, and models stay local
- End-to-end tuning: speculative decoding, concurrency, etc.
What It Can't Do
- •Windows is supported only through WSL, not natively
- •No fixed minimum hardware, but more memory allows larger models
- •Internet is needed only for initial installation and model download
- •The catalog is smaller than some rivals like Ollama, but you can add any GGUF model from Hugging Face
- •The project is evolving quickly; CLI and API may change between updates
Use Cases
- Developers who want to cut API costs and keep data private by using local models
- Users who already rely on agent harnesses and want to seamlessly switch them to a local backend
Magnitude is an open source inference server that runs the best local models for your hardware and plugs directly into the agent you already use (Pi, OpenCode, Hermes, OpenClaw, Codex, Claude Code, Oh My Pi, and Cline). Unlike generic model runners like Ollama, Magnitude does not rely on guesswork: it profiles your chip, memory, and bandwidth, then recommends models with estimated tokens/sec, auto-downloads and tunes them (speculative decoding, concurrency) for your specific machine. Models are loaded on demand and unloaded when idle or memory is tight, keeping prompts and files completely private and offline. Setup is agent-first: just tell your agent to install the CLI and run `magnitude docs onboarding`, and it walks you through everything. Magnitude supports macOS, Linux, and Windows via WSL, and is licensed under Apache 2.0.
Tags
Getting Started
Download installer
Click the button above to download the installer for your system
Install the software
Double-click the downloaded installer and follow the prompts
Install the Magnitude CLI: `npm i -g @magnitudedev/cli` (macOS/Linux, Windows via WSL)
Run `magnitude setup` to start the interactive setup, or paste the onboarding prompt into your agent to let it manage the process
Pick a recommended model; Magnitude downloads and tunes it, then switches your agent to the local backend
- Install the Magnitude CLI: `npm i -g @magnitudedev/cli` (macOS/Linux, Windows via WSL)
- Run `magnitude setup` to start the interactive setup, or paste the onboarding prompt into your agent to let it manage the process
- Pick a recommended model; Magnitude downloads and tunes it, then switches your agent to the local backend
SHA-256 checksum verified
Checksum extracted from GitHub official Release page
SHA256 Checksum
98f2a0afd5b1adf464fbfd093372899eed9cca48361514d73c49922fbf82ecd9This checksum is extracted from the GitHub Release page. Verify file integrity after download.
All SHA-256 checksums on this platform are extracted from the project's official GitHub Release page, without any modification. You can independently verify them on the GitHub Releases page.
Open Source Transparency
View GitHub SourceUninstall Info
Remove the CLI with `npm uninstall -g @magnitudedev/cli`; for a clean sweep delete the Magnitude data directory (usually ~/.magnitude) and restore any agent backend config you changed.
No Extra Dependencies
Ready to use after download. No additional runtime required.
Having issues? Check the FAQ below
5 FAQs
Similar Projects
unsloth
Unsloth is the first desktop app to run and train AI models locally. It supports a wide range of models including Qwen3.8, Kimi K3, MiniMax-H3, Gemma 4, DeepSeek-V4, and FLUX. Unlike most local model tools that focus only on inference, Unsloth adds a full training stack: fine-tune LLMs, diffusion, TTS, and embedding models 2× faster with 70% less VRAM. It also supports LoRA, QLoRA, RL, GRPO, DPO, and FP8. The app includes built-in Agents & Tools integration (Claude Code, Codex, MCP), private search and RAG, image/video diffusion, and audio support. With hardware support for CPU, NVIDIA, AMD, Intel, macOS, and multi-GPU setups, plus remote access via Cloudflare, Unsloth is a comprehensive local AI studio. It can export to GGUF, NVFP4, FP8 and offers an OpenAI-compatible API for easy deployment.
ragflow
RAGFlow is a leading open-source RAG engine that fuses deep document understanding with agentic workflows, offering precise, citation-grounded answers. Unlike LangChain's DIY approach, RAGFlow provides a turnkey, enterprise-ready platform with a visual interface and template-based chunking, reducing hallucinations and setup complexity.
ollama
Ollama lets you download, run, and manage large language models locally. One command, multiple platforms, endless possibilities.