Lemma launches an open-source workspace for teams and AI agents

Summary

Lemma is an open-source platform designed to let teams and AI agents work from the same permissioned context. Its core unit is a pod, which contains shared tables, searchable files, agents, workflows, functions, permissions, approvals, connectors, and one or more apps. Agents can run on schedules, webhooks, table events, chat, or API calls, while human approval steps can pause a workflow and resume later. The platform connects to Slack, Microsoft Teams, Telegram, WhatsApp, and email, with each surface reading and writing the same underlying records. Lemma also lets coding agents such as Claude Code, Codex, Cursor, OpenCode, or Antigravity generate the app, data model, agents, workflows, and permissions as files, then import and verify them through the CLI. Users can run the stack locally, deploy it on a server with Docker Compose, or use Lemma Cloud. Server-run agents support Lemma-managed models and OpenAI- or Anthropic-compatible providers, while local setups can connect to Ollama or LM Studio. The Python and TypeScript SDKs extend pod functionality to other applications. The core server components use AGPLv3, while the CLI, stack manager, SDKs, skills, and pod-bundle format use Apache-2.0. The documented desktop release targets macOS 14 or later on Apple silicon; no Windows installer is published.