Skip to main content
Magnitude is an open source inference server that profiles your hardware, recommends the best models for it, then downloads, tunes, and runs them. Use the Magnitude Harness or connect Pi, OpenCode, Hermes, or another supported harness.

Why Magnitude?

  • Knows your hardware: Profiles your chip, memory, and bandwidth.
  • Recommends what fits: Ranks exact model, quantization, and context combinations for your machine, with estimated tok/s.
  • No separate server: Includes its own inference engine, optimized for agent workloads.
  • Manages models for you: Loads models when your agent needs them and unloads them when idle or when memory is needed.
  • Tuned end to end: Configures acceleration, speculative decoding, context, and concurrency for your machine.
  • Set up in one command: Pick a recommended model and Magnitude connects and launches your harness.
  • Fully private and offline: Models, prompts, and files stay on your machine.
  • Free to run: No token costs, API keys, or rate limits.
  • Open source: Apache 2.0 licensed and yours to modify.
After setup, Magnitude runs in the background. Open your harness and work normally. Magnitude loads and serves the selected model as it is needed.