Your AI setup, in sync
on every machine. Nothing to run.

Every rule, skill, MCP server, subagent and memory in sync across your machines and harnesses. We run the server — you sign in and it works.

$ curl -fsSL mddb.cc/install | sh  ·  coming soon

mddb, the sync octopus

how it works

Three steps, then it disappears

Connect

Sign in and the dashboard hands you a device code with the mddb login line to paste. It enrolls the machine, scans it, finds Claude Code, Codex and Qwen Code, and imports rules, skills and MCP configs into your profile. You pick what syncs. Hosting mddb yourself? Compose brings up the same dashboard next to your server.

Work anywhere

Edit configs by hand, by CLI, or let your agent do it via MCP tools. Sync happens inside your harness sessions and on mddb sync: an edit you make while everything is closed lands the next time a session starts. Nothing runs in the background on your machine, and that is the point.

Never rebuild

New laptop, broken config, wiped disk: mddb restore brings back your exact environment, with a safety backup of whatever was there.

// devices

Devices

3 active

+ Add device
MachineAgentLast seenStatusProfileAttachments
Linux serverLinux 0.4.1just now onlinepaused: claude-code Research claude-code → Research
Work PCWindows 0.4.2just now online Content claude-code → Contentcodex → Engineering
MacBook PromacOS 0.4.2just now online Engineering claude-code → Engineeringcodex → Engineeringqwen → Research
schematic// every enrolled machine, its profile, its harnesses and what is paused
✓ Claude Code ✓ Codex ✓ Qwen Code OpenCode Gemini CLI iFlow CLI Kimi CLI Cursor + your harness: adapters are an open registry

what orbit syncs

Profiles, placement and per-harness control

Sync doesn't mean making every machine identical. Orbit keeps whole working roles behind one command, pins entities to the devices that need them, and lets a single harness override the machine it's on.

Profiles: a whole working role
in one command

A profile is the setup behind a kind of work, not a folder. Start from Engineering, Content or Research, or a Custom one you name yourself. mddb switch content swaps rules, skills, MCP servers and subagents across every harness on the machine at once, snapshotting what you had before the switch. Like branches, but for your AI setup.

MCP-first, no daemon

No background services. The sync engine lives inside an MCP server your harness starts with each session. Your agent can manage its own config by talking to it.

Restore on a clean machine

Sign in, then mddb restore: every rule, skill, MCP server and subagent comes back in the file each harness expects it in. A fresh install, your whole environment, nothing rewired by hand.

Placement, not flattening

Different release skills on work and home machines? Pin any skill to specific devices. Sync doesn't mean making every machine identical.

Per machine,
per harness

Six categories — instructions, skills, MCP servers, subagents, vault and memories — each set to sync, backup-only or off on every machine. The ones that map to harness files can be overridden for a single harness: skills reach Claude Code and stay out of Codex on the same laptop. Pause a whole machine or one harness with mddb pause --harness codex; resume catches its files up in one pass.

// profiles

Profiles

A profile is a line of work: engineering, content, research. Each carries its own rules, skills, agents and MCP servers.

+ New profile
  • Research

    Search and analysis agents, sources, note-taking skills.

    Machines: Linux server · 12m ago

  • Content

    Writing and editing prompts, style guides, publishing tools.

    Machines: Work PC · 12m ago

  • Engineering

    Coding agents, repo instructions, dev MCP servers.

    Machines: MacBook Pro · 12m ago

schematic// three lines of work, each carrying its own rules, skills and MCP servers

pricing

Pick your mass

Orbit plans run from a free tier to custom: zero setup, backups, history, support. Prefer your own server? Core is free and unlimited, forever.

Eight arms. Zero config drift.

We're building mddb in the open. Get on the list and be first to sync.

Join the waitlist