Glen

Glen is a shared learning layer for AI agents. It captures what an organization's agents and people learn as they work, then feeds the relevant context back into each new prompt, so that what one agent discovers is available to every other agent in the company. Capture is continuous and automatic. Glen collects agent sessions from Claude Code, Codex and Cursor, along with Slack, GitHub, tickets, calls and documents, and synthesizes them into one store. When someone starts a task, Glen finds the prior work that bears on it and injects that context, which the vendor says lets equivalent work finish faster and on fewer tokens. Glen exposes this through a single MCP tool that recalls the observations that matter and stores new ones in the same round trip. Other features include a company brain that can be queried from Slack, any MCP compatible agent or a web app, where code review and code search consult clones of the agents that made a change so users see the reasoning behind it. Skills distilled from a team's work surface automatically when someone tackles a related task, and transcripts, live session takeovers and shared artifacts work across harnesses and model providers. Glen suits engineering and knowledge work teams that run several coding or general purpose agents. Access is currently by waitlist, with teams onboarded in waves and a sign in for design partners. Pricing is not published.