One email each morning — yesterday's AI, sortedGet it in your inbox

METAL LAB

macro-inc/macro

3,806+2,557 this weekRust

An all-in-one workspace where email, chat, docs, tasks, and CRM are cross-linked and share one AI memory

Macro is an open-source workspace that merges email, messaging, docs, task management, CRM, and call logs into a single interface. Every item can be @mentioned and linked to every other item through a bidirectional graph, so AI agents can pull full team context and act on it. It's built with SolidJS and Rust and released fully under AGPLv3.

What it does

  1. Built because tools like Slack, Linear, Notion, HubSpot, and Superhuman don't talk to each other, leaving companies held together by MCP/Zapier glue and making operations hard to track as a whole
  2. Each feature ('block') is purpose-built for its job, but all blocks share one backend, so cross-references between a doc and a task, or a message and an email, are natively stored as a bidirectional graph
  3. A nightly cron job synthesizes team conversations, DMs, sent/received emails, and completed tasks into a single team-level memory, accessible via MCP or any AI model (OpenAI, Google, Anthropic)
  4. Coding agents like Claude Code or Codex can connect via MCP and use nearly 100% of what's available in Macro's UI, with no rate limits on MCP calls
  5. Made of 167 Rust libraries and 42 services on the backend plus a SolidJS client for web/desktop/mobile; built and dogfooded for two years by a ~15-person team in NYC and Toronto

Why it matters

For teams whose context gets scattered across disconnected SaaS tools and whose AI agents lose track of what's happening, Macro shows an approach built on one shared database and bidirectional links instead of stitched-together integrations. Being fully open source under AGPLv3 also matters practically, since it allows self-hosting and customization.

Terms in this repo

  • CRDT · a data structure that lets multiple people edit the same content simultaneously and merge changes automatically without conflicts, used for real-time collaborative docs
  • MCP · a connection standard that lets AI agents access external tools and data
  • bidirectional graph · a linking structure where if A references B, B is automatically recorded as referencing A back
  • hexagonal layout · a software design pattern that separates core logic from external integration code to ease maintenance

Repository description (English)

@-linked together with shared AI memory.

Open on GitHub

Trending repos

All repos →

Latest from METAL LAB