micro/mu

mu

A personal home server

Overview

A personal home server. One Go binary runs a set of everyday services — news, mail, markets, weather, search, images, video, blog, social, places, events — behind a Go Micro registry, with an LLM agent that calls them as tools. The same services are reachable as a web app, a REST API, an MCP server, an A2A endpoint, and a CLI.

Use it hosted at micro.mu, or self-host the single binary — same product either way. Open source, AGPL-3.0.

Features

• All services in one process. Each domain (news, markets, mail, weather, …) is a Go Micro service with typed handlers, registered in-process behind an in-memory registry. One binary, no external infrastructure; the same handlers can later be split across processes by swapping the registry.
• An agent over those services. An LLM — Claude, Atlas Cloud (DeepSeek), or a local Ollama / OpenAI-compatible endpoint — calls the services as tools, composes answers, and keeps per-user memory across sessions.
• A web UI that's a home screen. Cards render each service at a glance (headlines, prices, weather, unread mail); the agent sits inline to act on what you're looking at. Logged-out visitors get a public version with live public data.
• Several front doors to the same services. A REST API, an MCP server at /mcp, an A2A endpoint at /a2a, and a CLI where every tool is a subcommand. API and MCP callers can pay per request in USDC via x402.

Services

Each is reachable in the web app and directly over REST, MCP, A2A, or the CLI. The agent calls them as tools; each is also usable on its own.

• Agent — Ask anything. It calls news, markets, mail, weather, search and more, then synthesises an answer. Remembers your preferences.
• Chat — Conversational AI with session history
• News — Headlines from RSS feeds, chronological, with AI summaries
• Markets — Live crypto, futures, commodity, and currency prices
• Weather — Forecasts and conditions
• Mail — Private messaging and email (SMTP server with DKIM, inbound filtering)
• Social — Public discussion threads
• Blog — Microblogging with daily AI-generated digests
• Video — YouTube without ads, algorithms, or shorts
• Images — Generate images from a prompt, plus a daily nature / mindful image
• Search — Search the web without tracking, with a clean reader view
• Index — Search everything Mu holds for you: indexed news, blog, social and video, plus your own mail
• Places — Search places and nearby results with configured providers and open-data fallbacks
• Islam — Prayer times and a qibla compass, plus a Quran verse, hadith and reflection
• Events — Schedule reminders by asking the agent; delivered to your channels, with a calendar invite
• Apps — Build and use small, useful tools — pin any app to the top of your home screen. Apps can call every service above through the SDK, so a new service is available to them the moment it registers
• Stream — Public event feed for agents and tools to subscribe to

Building on it

Every capability is a service behind one registry, and that registry is what everything else reads. Register a service and it becomes available, in the same moment and with no extra wiring, to:

• the agent, as a tool it can call;
• custom agents, in the tool picker at /agent/new;
• the MCP server, for Claude Desktop, Cursor or any MCP client;
• the REST API and the CLI, where each tool is a subcommand;
• apps, through mu.service(name, method, args) in the SDK.

So extending Mu is adding an element, not wiring N integrations. Apps and agents are the two things built on the services — an agent is a system prompt plus a chosen subset of tools; an app is a page with the SDK. Both get new capabilities for free as the service set grows.

// inside an app — any registered service, no SDK change needed
const { times } = await mu.service('islam', 'prayer', { lat, lon, tz })
const { summary } = await mu.service('weather', 'forecast', { lat, lon })

Identity is bound server-side: an app never names whose data it wants, and account-scoped services (mail, index, images, events) require a signed-in caller.

Accounts & sign-in

Sign in to the web app with a username and password, a passkey (WebAuthn), or Google. Already have an account? Link Google to it from Account settings and use Google sign-in from then on. For the API and CLI, generate a Personal Access Token at /token.

Passkeys work out of the box. To enable Google sign-in when self-hosting, set GOOGLE_CLIENT_ID and GOOGLE_CLIENT_SECRET (and optionally GOOGLE_REDIRECT_URI, which defaults to <your-origin>/oauth2/callback) from /admin/env or the environment.

For Agents

Because every capability is a service, it's reachable however you like. Mu exposes a REST API and an MCP server at /mcp, so AI agents and tools can connect directly.

{
  "mcpServers": {
    "mu": {
      "url": "https://micro.mu/mcp"
    }
  }
}

30+ tools — news, search, weather, places, video, email, markets, images, events — accessible via MCP. AI agents can pay per-request with USDC through the x402 protocol. No API keys. No accounts. Just call and pay. First 10 calls per wallet are free.

See MCP docs

CLI

Every MCP tool is also available as a mu subcommand. The same binary runs the server (mu --serve) and the CLI.

mu news                                 # latest news feed
mu news_search "ai safety"              # search news
mu chat "hello"                         # chat with the AI
mu agent "what is the btc price?"       # run the full agent
mu search_web "claude code"             # search the web
mu weather_forecast --lat 51.5 --lon -0.12
mu me                                   # your account
mu help                                 # full tool list

The CLI is registry-driven — every tool added to the MCP server automatically becomes a CLI command.

Authentication

mu login                  # opens /token in your browser, paste the PAT back
mu config set token xxx   # or set it directly
export MU_TOKEN=xxx       # or use the environment

See CLI docs for more.

Discord & Telegram

Talk to the AI agent from Discord or Telegram. Ask questions, check markets, get news — all from chat.

Join the Discord

Discord slash commands: /agent, /news, /markets, /weather, /mail, /social, /blog, /video, /search, /apps, /balance, /usage. Telegram commands: /agent, /ask, /news, /markets, /weather, /usage.

See Discord docs and Telegram docs for setup.

Self-hosting

One-command install

curl -fsSL https://raw.githubusercontent.com/micro/mu/main/install.sh | sh

Docker

git clone https://github.com/micro/mu && cd mu
docker compose up

From source

git clone https://github.com/micro/mu
cd mu && go install
mu --serve

First-run setup

Open http://localhost:8080 and Mu walks you through a one-time setup: create your admin account and pick an AI provider (Claude, Atlas Cloud, or a local Ollama / OpenAI-compatible endpoint). That's enough to have a working agent.

Prefer the terminal? Configure the provider headless, then start the server:

mu setup        # pick a provider, paste a key
mu --serve      # first account you create becomes admin

Or set everything by hand if you'd rather:

export [email protected]          # who's admin (else: first account)
export ATLAS_API_KEY=xxx              # or ANTHROPIC_API_KEY, or OPENAI_BASE_URL
mu --serve

Once you're admin, every other key (YouTube, Brave search, weather, mail/DKIM, Google sign-in…) is configurable from /admin/env in the browser.

See Installation guide for full setup.

Configuration

Customise feeds, prompts, and cards by editing JSON files:

• news/feeds.json — RSS news feeds
• chat/prompts.json — Chat topics
• home/cards.json — Home screen cards
• video/channels.json — YouTube channels
• places/locations.json — Saved locations

See Environment Variables for all options.

Documentation

Full docs in the docs folder.

License

AGPL-3.0 — use, modify, distribute. If you run a modified version as a service, share the source.