Automatic 1.0: Your Agents, Finally in Sync
On this page
Automatic 1.0 is here. After months of early releases, feedback, and iteration, Automatic is now a stable, production-ready desktop hub for managing your AI coding agents.
If you've been using AI coding agents seriously — Claude Code, Cursor, Codex, GitHub Copilot, or any of the others — you've probably hit the same wall: getting great results requires carefully built context, but that context is trapped in individual projects and individual tools. You end up rebuilding it from scratch, or worse, you get inconsistent results because every tool is configured differently.
Automatic fixes that. One app, one source of truth, synced everywhere.
What Automatic does
Automatic sits between you and your agents. You configure your skills, MCP servers, instructions, and rules in Automatic, and it syncs the right configuration into the right format for each supported tool.
Your agents don't know Automatic exists. They just read their standard config files as usual. But those files are now consistent, complete, and always up to date.
Skills that follow you between projects
Skills are reusable bundles of specialised knowledge and workflows that make your agents better at specific tasks. Automatic lets you install, organise, and assign skills to any project.
19 skills ship built-in, covering common workflows like code review, testing, and architecture design. The marketplace offers community-contributed skills for more specialised needs — from Terraform infrastructure to Laravel development to lead magnet copywriting.
The key benefit: build context once, reuse it everywhere. A skill you install for one project is immediately available to any other.
MCP servers, centrally managed
Model Context Protocol servers give your agents access to external tools and services — Linear for issue tracking, Sentry for error monitoring, GitHub for repository operations, and dozens more.
Automatic manages MCP server configuration centrally. Register a server once, assign it to the projects that need it, and Automatic handles the per-tool config format differences. No more maintaining separate MCP configs for Claude Code, Cursor, and every other tool.
Automatic also exposes its own MCP server, so agents can pull project context, stored memories, and credentials directly during a task.
Instructions and rules that stay in sync
Every AI coding agent has its own instruction format — .cursorrules for Cursor, CLAUDE.md for Claude Code, AGENTS.md for Codex, and so on. Automatic lets you write unified project instructions and translates them into the right format for each agent.
Rules are composable and reusable. Define a rule once — coding standards, architectural patterns, security guidelines — and apply it across any combination of projects and tools. When you update a rule, every project that uses it stays current.
For agents that lack certain features natively, Automatic polyfills the gap. Commands in Codex, split rules outside Claude Code — Automatic bridges the differences so you get a consistent experience regardless of which tool you're using.
Agent memory that persists
Automatic provides project-scoped memory stores where agents can save decisions, conventions, discovered gotchas, and working context. Memory persists across sessions and across tools — an insight captured by Claude Code is available to Cursor on the next task.
Agents access memory over MCP, so there's nothing to configure on the agent side. They read and write naturally as part of their workflow, and the knowledge accumulates over time.
Drift detection
When configurations diverge — a teammate edits a file directly, an agent overwrites a setting, or a sync falls behind — Automatic flags it. Drift detection monitors your projects and alerts you when things are out of alignment, so small inconsistencies don't compound into hard-to-debug problems.
Credential management done right
API keys and tokens are stored in your system keychain, not in config files. Agents retrieve credentials over MCP when they need them, which means sensitive values never end up committed to version control or scattered across dotfiles.
Project templates
Bundle a complete project setup — skills, MCP servers, instructions, and rules — into a reusable template. When you start a new project or onboard a teammate, apply a template and everything is configured in seconds.
16 supported agents
Automatic syncs with Claude Code, Cursor, GitHub Copilot, Codex, Cline, Zed, Warp, Gemini CLI, Kilo Code, Kiro, Antigravity, Goose, OpenCode, and more. The list continues to grow with each release.
Not everyone on your team needs Automatic. Because it writes standard config files, teammates who don't use Automatic still benefit from the configurations it manages — they just read the files their tools already expect.
Available now
Automatic 1.0 is available for macOS, Windows, and Linux. It's free for internal use with up to 10 concurrent users, source-available under the Business Source License, and converts to MIT in 2030.
Download Automatic and get your agents in sync.
Related reading
Automatic 0.8.0: AI-Generated Instructions and Recommendations
Version 0.8.0 brings AI-powered instruction generation, skill and MCP recommendations, project context via a dedicated MCP tool, and a major overhaul of how rules and context are managed across your projects.
Introducing Automatic: One Config for Every Agent
We built Automatic because managing AI agents across tools had become a nightmare. Here's why we did it, and what we built.
Add Automatic Support to Your Repository
Ship skills, MCP servers, rules, and more directly from your git repository. One manifest, one-click install, every AI agent supported.