Syzygy

Research lab building more capable models with less memory, compute, and energy, focused on efficient, locally controlled intelligence.

Past recognitions

About Syzygy

Research lab building more capable models with less memory, compute, and energy, focused on efficient, locally controlled intelligence.

Products

  • Mach Studio

    Coding Assistants & AgentsFree plan

    View product

    Desktop application for running Claude Code, Codex, and OpenCode with local models, subscriptions, cloud models, integrations, automations, and remote control.

  • Mach 2 Additive Medium

    Foundation & Language ModelsFree plan

    View product

    AI model designed for efficient inference, with 125.7 billion text parameters, 1.70 bits per weight, and 26.7 GB on disk.

  • Mach 1 Additive Small

    Foundation & Language ModelsFree plan

    View product

    35B mixture-of-experts AI model supported on Macs with 16 GB or more of memory.

  • Syzygy Inference

    Model Hosting & InferenceFree plan

    View product

    Syzygy-hosted inference service listed as an early-access offering.

Security and compliance

As reported by Syzygy. Ask for audit reports before you buy.

  • Local-first architecture with on-device storage for chats, prompts, generated artifacts, files, and agent memory
  • On-device model inference and chat content remain on the user's Mac
  • TLS encryption in transit between the app, Syzygy services, and providers
  • Provider-supplied protections for persistent databases and object storage
  • Production credential restrictions and provider identity and access controls
  • OAuth authorization for connected third-party tools where available
  • Data minimization and content-free allowlisted product analytics outside Europe; Mach Studio usage analytics are not collected from Europe

Is this your company?

Claim the profile to keep facts current, answer buyers and submit evidence for the next ranking review.

Claim Syzygy