GitHub renders `<org>/.github/profile/README.md` as the org's public landing page. Mirroring the same path here on Gitea so once GitHub access is restored and repo-sync pushes this content over, it lands in the right place automatically — no path translation step. ## Adds - `profile/README.md` — comprehensive org-level intro (different from molecule-core's product-deep README): pitch in 1 paragraph, "what we ship" in 4 bullets, "where to start" routing table, repos-in-this-org index split by purpose (product / adapters / marketing), license + community footer. - `profile/assets/molecule-icon.svg` — same SVG that landed on molecule-core's README in PR #5; light/dark adaptive via prefers-color-scheme styles. - `profile/assets/molecule-logo.svg` — wordmark variant. ## Updates - `README.md` (repo root, not user-visible on the org page) — documents the layout convention + lists the other cross-org defaults GitHub picks up from `<org>/.github/` (CONTRIBUTING.md, SECURITY.md, SUPPORT.md, issue/PR templates, workflow templates) so a future contributor knows where to add things. ## Why this content shape GitHub org-profile READMEs are user-visible on the org's public page. Standard wisdom for that surface: - short — 1-2 screens, not the deep product pitch (which lives at molecule-core) - routes the reader to the right repo for what they actually want (product / SaaS / Claude integration / docs / new runtime) - lists the major repos with one-line descriptions so a visitor can navigate without 30 tabs - names the license model (BSL 1.1 → Apache 2.0 on 2029-01-01) The "Where to start" table is the load-bearing piece — the rest is context. If a visitor only reads the table, they still know where to go. ## Out of scope (parked) - CONTRIBUTING.md, SECURITY.md, SUPPORT.md — would normally live here too. Skipping for now; current per-repo guidance is fine. File when the org has enough cross-cutting policy to warrant it. - GitHub Actions workflow templates — not relevant while we're on Gitea + operator-host deploys. 🤖 Generated with [Claude Code](https://claude.com/claude-code) |
||
|---|---|---|
| .. | ||
| assets | ||
| README.md | ||
Molecule AI
The Org-Native Control Plane for Heterogeneous AI Agent Teams
Molecule AI is the operating system for AI agent organizations. One control plane, eight runtimes, organization-aware memory, real ops surfaces — built for teams that want to move past the demo.
What we ship
- Org-native workspace abstraction — every node is a role with persistent identity, memory boundary, hierarchy position, and A2A interface. Workspaces grow into teams without breaking external integrations.
- Eight production agent runtimes behind one workspace contract — Claude Code, Hermes 4, Gemini CLI, LangGraph, DeepAgents, CrewAI, AutoGen, OpenClaw. Mix them across teams without giving up shared governance.
- Memory v2 with pgvector semantic recall, scoped to organizational boundaries (
LOCAL/TEAM/GLOBAL). Memories promote into hot-loadable skills as workflows mature. - A real operational plane — Canvas v4 (warm-paper themed), live activity streams, restart/pause/resume, terminal + files into running workspaces, Langfuse traces, the whole picture.
Where to start
| If you want to … | Go to |
|---|---|
| Read the pitch + run it locally | molecule-core — Apache 2.0-converting BSL 1.1, Quick Start in 5 commands |
| Try the hosted SaaS | moleculesai.app — multi-tenant, EC2 + Cloudflare Tunnels, WorkOS auth, Stripe billing |
| Bridge a Claude Code session into a Molecule workspace | molecule-mcp-claude-channel — /plugin marketplace add Molecule-AI/molecule-mcp-claude-channel |
| Browse the architecture | docs/architecture/architecture.md |
| See the memory model | docs/architecture/memory.md |
| Add a new runtime / template | molecule-ai-workspace-runtime (mirror of workspace/) + the molecule-ai-workspace-template-* repos |
Repos in this org
Product
molecule-core— the platform. Go control plane + Next.js Canvas + Python workspace runtime. Open source under BSL 1.1 (converts to Apache 2.0 on 2029-01-01).molecule-controlplane— SaaS multi-tenant orchestration on top ofmolecule-core. Private; backs moleculesai.app.
Adapters & integrations
molecule-mcp-claude-channel— Claude Code MCP plugin: bridges Molecule A2A traffic into a local Claude Code session.molecule-ai-workspace-runtime— Python adapter runtime, mirrored frommolecule-core/workspace/. Published as a PyPI wheel.molecule-ai-workspace-template-*— one repo per supported runtime template (claude-code, hermes, langgraph, …).
Marketing & community
landingpage— moleculesai.app source.molecule-ai-status— public status page.
License & community
Most user-facing code is open source under the Business Source License 1.1 (converts to Apache 2.0 on 2029-01-01). Personal, internal, and non-commercial use is permitted without restriction; commercial use that competes with the hosted product is what the BSL gates.
Issues, PRs, and design discussions land on the individual repos above. The platform repo's Docs Home is the canonical entry point for engineering depth.
Org profile last refreshed 2026-05-06.