A coding agent is an AI assistant that runs on your computer. It is how apps get built on Syntica: you talk to it, and it writes, changes and deploys the app. One command sets everything up:
$ npx @syntica/mcp setupRestart your agent afterwards and you're connected. The command finds the agents already on your machine and connects each of them: Claude Code, Codex, Cursor, Windsurf, Gemini CLI and Copilot CLI. The agent pages cover each one.
/syntica:onboard.Open your agent in an empty folder and describe the app in your own words, the way you would brief a developer:
The agent asks clarifying questions, then creates the app: screens, database, logic. Your job is the functional one: check that it matches what you meant, and keep refining in plain language. You never have to read the code.
In Claude Code, run /syntica:onboard. It sets everything up, including your account: you answer a few questions (name, email, organisation name) and confirm with a one-time code. No registration form, no password. Then run /syntica:deploy and you get a live address back.
On the other agents, just say what you want: "set me up on Syntica and deploy this app." The agent walks the same path with the same tools, including signing you up.
Every app gets separate development, staging and production environments. It's the same deploy step, pointed at a different one.
Everything the agent does shows up at app.syntica.io. Sign in with the same email and a one-time code: your apps, every deployment and its status, logs, and a record of every change. The dashboard is where you manage and check. Building stays in the conversation with your agent.