DEVOURED
Lemma (GitHub Repo)
Lemma provides an open-source, multi-agent workspace that treats state as shared, permissioned data across teams, surfaces, and autonomous agents.
What: Lemma creates 'pods'—self-contained environments holding shared tables, files, workflows, and permissions—that allow human teams and AI agents (Claude, Codex) to collaborate on shared records and logic.
Why it matters: By standardizing the 'harness' (state, permissions, and tool access) around agents, Lemma moves beyond single-user agent sessions toward persistent, team-wide business systems.
Takeaway: Install the Lemma CLI using `uv tool install lemma-terminal` to start building pod-based workflows locally.
Deep dive
- Pods: Encapsulated environments for shared state, workflows, and permissions.
- Agent Host: Bridges local coding agents (Claude Code, Cursor) into the pod ecosystem.
- Surfaces: Integration layer connecting Slack, Telegram, WhatsApp, and email to pod workflows.
- Dual-licensing: AGPLv3 for the core backend/frontend; Apache-2.0 for client-side SDKs and CLI.
- Portability: The same code runs on a laptop, local VM, or the managed Lemma Cloud.
Decoder
- Pod: The basic unit of organization in Lemma, containing data (tables), memory (files), agents, workflows, and UI definitions.
- Harness: The surrounding infrastructure (memory, tool access, state) that allows an AI agent to execute tasks within a defined environment.
Original article
Full article content is not available for inline reading.
Explore all 57 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi