YC Open-Sources Internal AI Infra 'QM', Providing Enterprise Agent Blueprint
eyishazyer · x · 2026-08-02
Y Combinator has open-sourced its internal AI infrastructure, QM, under the MIT license on GitHub. Originating from real organizational use at YC, it provides every employee and project with an AI workspace featuring isolated memory, permissions, and persistent sandboxes, with seamless syncing between Slack and web apps.
Key Features:
- Model-Agnostic: Supports multiple models like Claude Code and Codex. Models can be swapped without losing the company's memory, permissions, or workflows.
- Continuous Execution: Agents can search internal info, monitor CI, run tests, and build web apps. They continue working in the background via crons even when browsers are closed.
- Shared Skills: Workflows created by one employee can be adopted by the team or imported as skill packs from Git repositories.
- Tech Stack: The core is powered by TypeScript and Fastify, with Postgres handling memory and session storage.
Security Disclaimer: YC explicitly labels QM as early, experimental software. Its documentation openly lists prompt-injection gaps and bypassable controls, meaning it is not yet enterprise-hardened.
Related event: Y Combinator open-sources internal multi-agent framework QM under MIT(15 posts)→
More from coding & agent
- Connected steps are not a workflow: nobody designs the handoff — usehive · 2026-10-16
- Making Blender animations with Opus, Astra and Argon: how you prompt still matters — chongdashu · 2026-10-16
- Running PolyStrata + Cline coding agent on a $0-fancy laptop at ~90 tok/s — rpwoerk · 2026-10-16
- How do you attribute agent gains: model training or infrastructure? — jrxna · 2026-10-16
- Feed Claude your books, blogs and top engineers' code for claude.md — 'it now writes better code than me' — blaizedsouza · 2026-10-16
- Open-source /label plugin auto-names and colors Claude Code sessions in one step — hermanschutte · 2026-10-16