Meland Labs builds open-source AI coworkers and the context runtimes that power them — connecting the tools you already use and surfacing only what needs your decision or action.

OpenLoomi is an open-source AI coworker built to protect your attention. It connects the tools you use and, with your permission, understands what's on your screen — bringing together the context behind your people, projects, and decisions, then surfacing only what needs your decision or action.

An always-on desktop attention agent (Loomi) watches the door so you don't have to, and turns the day's scattered signals into decision cards you can approve in one tap. Use it standalone, or plug any Agent framework into the same resident desktop — Claude Code, Codex, OpenCode, Hermes, and OpenClaw all work.

• 🐾 Attention Agent — An always-on desktop companion (Loomi) that surfaces pre-decided reminders — 9 AM to-do, 6 PM recap, overdue replies — as small bubbles, without fighting for your focus.
• 🧠 Holistic Context — Short → mid → long-term memory that grows on its own — visible, auditable, and always remembering your people, projects, and decisions across months.
• 🔌 Platform Connectors — Background sync loop pulls commits, issues, emails, and docs into your context graph. Chat with AI directly inside Telegram, WhatsApp, iMessage, QQ, or Lark/Feishu.
• ⏰ Proactive Tasks — Schedule recurring work — daily digests, weekly reports, reminders — that run automatically on your desktop.
• 🖥️ Local-First Security — Native apps for Windows, macOS, and Linux with AES-256 encryption, no data leaves your machine, auditable access logs.
• 🧩 Any Agent Integration — Context, memory, connectors, attention agent, and Loop engine are all delivered as open-source Skills and Plugins — plug into Claude Code, Codex, OpenCode, Hermes, or OpenClaw.

OpenContext is the agentic context runtime that powers OpenLoomi — and is open-source so any team can embed it. It bundles a temporal context graph, memory API, retrieval primitives, multi-platform integration mesh, and a deterministic loop engine behind a single pnpm add @melandlabs/opencontext, so you can ship a durable, auditable agent in days instead of quarters.

• 🧠 Temporal Context Graph — facts carry valid_from / valid_until; supersession, contradiction, and merge are first-class edges with append-only corrections.
• 🔌 Integration Mesh — uniform IntegrationRecord across Gmail, Slack, Telegram, Linear, Jira, iMessage, Feishu, and Weixin.
• ⏰ Deterministic Loop Engine — the scheduler wakes up, decides if there's real work, then calls the agent — never the other way around.
• 🔍 Retrieval Primitives — chunking, embeddings, parsers, sqlite-vec / pgvector / Chroma adapters.
• 🤖 Agent Runtime — AI SDK wrappers, sandbox providers, MCP server, and memory consolidation.
• 🪶 Library-First API — add one dependency, then embed it, run the HTTP daemon, or wire it as an MCP server into Claude Desktop / Cursor.
• 🛡️ Audit & Encrypted Storage — JSONL audit logs, Fernet encryption, URL allow/blocklist by default.

• Stop dropping commitments. Overdue replies, creeping deadlines, "I'll send that follow-up Friday" — Loomi nudges you with a small bubble at the right moment, and you can teach it the signals and decisions to watch for.
• Find anything from your work life in seconds. "What did I decide about pricing last quarter?", "Who owns design on Acorn?", "What was I working on before vacation?" — memory spans every tool and channel, no more digging through Slack, Gmail, or Notion.
• Morning brief at 9 AM, end-of-day recap at 6 PM. Today's to-do slides into view every morning, today's done ships every evening — the key info lives in one place instead of nine apps.
• Ask AI for work help right inside your chat. Draft a reply, summarize a long thread, schedule a follow-up — directly in Telegram, WhatsApp, iMessage, QQ, or Lark/Feishu.

• 🌐 openloomi.ai
• 📖 Documentation
• 🐛 Issues
• 💬 Discussions
• 👾 Discord
• 🐦 X
• 🧠 OpenContext on GitHub