Documentation

1. Quickstart (Installation)

RingDeveloper is a Model Context Protocol (MCP) server. There are two ways to connect it — pick whichever your editor supports. Either way you'll need a verified phone number so the agent can call you.

Option A — OAuth (recommended)

For OAuth-capable clients (Claude Code, Claude Desktop, Cursor, VS Code, Zed), just add the hosted MCP server URL — no API key:

Your editor opens a browser to sign in to RingDeveloper and approve access, registering itself automatically (Dynamic Client Registration). Manage authorized apps anytime under Connected apps — where you can also create a static client (client ID + secret) for tools that need one.

Option B — API key (CLI / any MCP client)

For CLI agents, cloud platforms, or any client without OAuth, use a key. The interactive setup wizard configures your editor automatically:

Run this in your codebase's terminal. It asks for your API key and verified phone number, then generates a .mcp.json in your project.

2. API Keys & Verification

3. Manual IDE Configuration

If your IDE requires manual configuration in its UI settings instead of reading .mcp.json, add the server using this snippet. OAuth-capable clients(Claude Code, Cursor, VS Code, Zed, Claude Desktop) don't need this — use Option A (the MCP URL + browser sign-in) instead.

"mcpServers": {
  "ringdev": {
    "command": "npx",
    "args": ["-y", "@ringdev/mcp"],
    "env": {
      "RINGDEV_API_KEY": "rdk_live_...",
      "RINGDEV_PHONE": "+1234567890"
    }
  }
}

4. Tell your agent how to use it

Connecting makes the tools available — this prompt makes your agent actually call them at the right moments. The setupwizard writes it into your IDE's rules file automatically; paste it yourself if you configured the server by hand.

RingDev — you can phone the developer through the RingDev MCP tools. Never wait silently, and never guess on a decision that's the human's to make.

Tools:
- ring_developer(question, context, urgency) — ask ONE specific, blocking question or request an approval. The developer answers by voice; their reply is your next instruction.
- notify_completion(summary, next_step_options) — when a unit of work is done, report it and offer next steps; the spoken choice is your next instruction.
- check_inbound_requests() — pick up coding tasks the developer queued by phone.
- complete_inbound_request(session_id, summary) — mark a voice-initiated task complete after you finish it.

Call ring_developer / notify_completion whenever you would otherwise pause for the human: an approval or permission prompt, a plan or "proceed?" sign-off, a "continue to iterate?" confirmation, a clarifying question, an ambiguous choice, a risky or irreversible action, a failing build you can't decide on, or handing control back at end of turn. Ask one clear question with enough context to answer without seeing your screen.

Call check_inbound_requests at the start of every session (launch, resume, or first message), at the start of each turn, and after finishing a task — then do whatever was queued. After finishing a voice-initiated task, call complete_inbound_request with its session_id so it leaves the developer's active dashboard.

If a RingDev tool errors or times out, use your best judgment and keep going — never block. Redact secrets (keys, tokens, passwords) to [redacted] before sending.