Magnitude AI Inc.
Magnitude AI Inc. builds an open-source local inference engine/server that profiles a user's hardware, recommends models that fit, then downloads, tunes, and runs them locally. It plugs into existing coding/AI-agent harnesses so prompts, files, and models can remain private and offline.
At a Glance
- Individual developers
- Coding-agent users
- Open-source and infrastructure developers
- Privacy-sensitive users and teams
- +1 more
AI Tools by Magnitude AI Inc.
(1)
Magnitude
Local AI Inference Engine
Discussions
No discussions yet
Be the first to start a discussion about Magnitude AI Inc.
Latest News
GitHub released @magnitudedev/cli@0.0.15-alpha.1 with a macOS AssociatedBundleIdentifiers fix so the login item shows the correct icon and name.
GitHub released @magnitudedev/cli@0.0.15-alpha.0 with Apple signing.
Released @magnitudedev/cli 0.0.14, with updated Pi plugin UI and downloadable macOS/Linux assets.
Magnitude posted a Pi-user-focused product update and a local-model radar-chart update on its official X account.
Products & Services
Open-source local inference server that profiles chip, memory, and bandwidth; ranks models by speed, accuracy, intelligence, memory, and estimated tok/s; downloads and tunes a selected model; and loads/unloads models on demand.
Global CLI installed with npm that guides hardware profiling, model recommendation, download, tuning, and connection to an agent harness; current stable package version observed was 0.0.14.
Optional built-in agent harness available alongside integrations with external coding-agent harnesses.
Market Position
Magnitude positions itself against hosted inference and simpler local-model setup tools such as Ollama. Its differentiation is an agent-first workflow that profiles the user's hardware, recommends models and quantizations that fit, estimates performance, tunes the stack, and manages model lifecycle automatically, while integrating with multiple coding-agent harnesses.
Leadership
Founders
Tom Greenwald
Founder/CEO of Magnitude; studied computer science at Northeastern and ran special projects as a product manager at SimpliSafe. Previously co-founded Sidekick with Anders Lie.
Anders Lie
Founder/CTO of Magnitude; studied computer science and mathematics at Iowa State, and was a software engineer at AWS working on no-code automation platforms. Previously co-founded Sidekick with Tom Greenwald.
Executive Team
Tom Greenwald
Founder/CEO
Northeastern computer-science graduate; former special-projects product manager at SimpliSafe.
Anders Lie
Founder/CTO
Iowa State computer-science and mathematics graduate; former AWS software engineer focused on no-code automation platforms.
Founding Story
Magnitude emerged from the founders' work on Sidekick, an automation product they launched through YC. Its initial vision was to make capable AI agents genuinely local: instead of sending prompts, files, and secrets to hosted model providers, run the inference stack on the user's own computer, with hardware-aware model selection and one-command setup.
Business Model
Revenue Model
The public Magnitude product is free to run and open source; the company publicly emphasizes no token costs, API keys, or rate limits. The available sources do not document a paid pricing model or reported revenue.
Pricing Tiers
No token costs, API keys, or rate limits; Apache 2.0 licensed.
Target Markets
- Individual developers
- Coding-agent users
- Open-source and infrastructure developers
- Privacy-sensitive users and teams
- Organizations running AI workloads on local hardware
- Running coding agents on private local models
- Analyzing sensitive data and private notes without sending them to cloud model providers
- Reviewing code, logs, and local files
- Building documents or slides with an agent
- Offline or cost-free local inference
- Using local models on developer laptops and workstations