--- title: "Introducing Trophy's MCP Servers: Build and Run Gamification From Your AI Agent" canonical_url: "https://trophy.so/blog/introducing-trophys-mcp-servers" description: "Trophy now has two MCP servers that let AI agents like Claude, ChatGPT and Cursor build gamification into your app for you. The Trophy Docs MCP server gives agents live access to Trophy's documentation. The Trophy Account MCP server lets them set up and manage your Trophy account from chat." last_updated: "2026-09-30T14:23:06.000+00:00" --- # Introducing Trophy's MCP Servers: Build and Run Gamification From Your AI Agent **Trophy now has two MCP servers that let AI agents like Claude, ChatGPT and Cursor build gamification into your app for you.** The Trophy Docs MCP server gives agents live access to Trophy's documentation. The Trophy Account MCP server lets them set up and manage your Trophy account from chat. Trophy is gamification infrastructure for consumer apps. Product teams use its APIs and SDKs to add achievements, streaks, points, leaderboards and personalised notifications without building the backend systems themselves. With the MCP servers, much of that work now happens inside the AI tools your team already uses. ## TL;DR - **Two servers:** Trophy Docs MCP (`https://docs.trophy.so/mcp`) for knowledge, Trophy Account MCP (`https://mcp.trophy.so/mcp`) for action. - **Works where you work:** Claude, Claude Code, Cursor, ChatGPT, VS Code and other MCP clients. - **No API keys:** the Account server uses OAuth sign-in, scoped to one Trophy organisation. - **Built to be safe:** changes target a named environment, and agents can't delete your configuration. - **Get started:** [MCP documentation](https://docs.trophy.so/mcp/overview). ## Why we built it Adding gamification with Trophy has always involved two kinds of work. The first is code: sending events from your app to the [Trophy API](https://trophy.so/developers). The second is configuration: deciding which actions count, how many points they're worth, when a leaderboard resets and what unlocks an achievement. Code was already moving into AI agents. Most of the teams we work with design and ship product inside Claude Code, Cursor and similar tools. Configuration, though, still meant leaving that flow to click around a dashboard, then explaining the result back to the agent. We wanted to close that gap. Over the past few months we've been building towards it: an Admin API for account configuration in April, and public OpenAPI specs in May so coding agents always have current information. The MCP servers are the final step. An agent can now work out what gamification your product needs, map it to Trophy's features, configure your account and write the integration code, all in one conversation. ## What the MCP servers do ### Trophy Docs MCP: live documentation for your agent The Docs server is read-only. It lets an agent search and read Trophy's current documentation and API specs, instead of relying on whatever it remembers from training. That means fewer invented endpoints and more integrations that work first time. ### Trophy Account MCP: the action layer The Account server lets an agent read and manage your Trophy account: metrics, points systems, achievements, leaderboards, streaks and settings. You describe what you want in plain language and the agent does the rest: - _"Create an XP system called 'Kudos' that grants 10 points for each workout completed."_ - _"Get the top 20 rankings for leaderboard weekly-workouts this week."_ - _"Grant two streak freezes to user\_123."_ The two servers are designed to work together. The agent learns how a feature works from the docs, checks what already exists in your account, creates only what's missing, then writes the matching code in your app. ## Why use it **Integrate faster.** A coding agent that can read the docs and configure your account can take you from idea to a working streak or points system in a single session. **Get answers from your data without a dashboard.** Ask for achievement completion rates, leaderboard rankings or how points are distributed across paying users, and get the answer in chat. **Hand off routine operations.** Restoring a user's broken streak, granting freezes or scheduling a weekend XP boost become one-line requests, so support and growth teams don't need an engineer. **Bring Trophy into your own agent workflows.** Because it speaks the open Model Context Protocol, Trophy can sit alongside the other tools your agents use. ## Built with guardrails Letting an agent change live configuration needs care, so we made some deliberate choices: - **Scoped access.** You sign in with OAuth and authorise a single Trophy organisation. There are no keys to leak. - **Explicit environments.** Every change targets a named environment, so you can iterate in staging before touching production. - **No deletions.** Agents can create and update configuration but can't delete it. The only exception is removing a temporary streak pause. - **Chat, not cron jobs.** The MCP servers are for interactive work with an agent. Production integrations should keep using the Trophy API directly. ## It gets better as you use it Every call to the Account server includes a short note from the agent on why it's making the call. When an agent needs something the server can't do yet, it can say so and describe the tool it wishes existed. Those requests come straight back to us, so we learn what to build next from the agents actually using it. ## Get started 1. [Create a Trophy account](https://app.trophy.so/sign-up) if you don't have one. 2. Add `https://docs.trophy.so/mcp` to your MCP client as **Trophy Docs**. 3. Add `https://mcp.trophy.so/mcp` as **Trophy Account**, then sign in and choose your organisation. 4. Ask your agent to build something. Setup guides for each client are in the [installation docs](https://docs.trophy.so/mcp/account-mcp-server/installation), and our [best-practice guide](https://docs.trophy.so/mcp/best-practice) covers how to get the most from both servers together. ## FAQ ### What is the Trophy MCP server? Trophy has two remote MCP servers. The Docs MCP server gives AI agents read-only access to Trophy's documentation. The Account MCP server lets agents read and manage gamification configuration in a Trophy account. ### Which AI tools work with Trophy's MCP servers? Any MCP client, including Claude, Claude Code, Cursor, ChatGPT and VS Code. ### Do I need an API key? No. The Account MCP server uses OAuth, so you sign in with your Trophy account. The Docs MCP server doesn't even need a Trophy account, since it only reads public documentation. ### Is it safe to let an agent change my Trophy account? Access is limited to the organisation you authorise, changes target a specific environment, and agents cannot delete configuration. We recommend trying changes in a staging environment first, promoting to production once ready. ### How much does it cost? The MCP servers are available on all Trophy accounts at no extra cost.