Gamebeast MCP server
Your AI agents, working in your analytics
Connect Claude, Cursor, or Codex to Gamebeast. Over MCP, your agent can read your data and, with write access, change configurations and run experiments, each within your permissions. Free on every plan.
What your agent can do
Read everything
Analytics (insights, funnels, retention, heatmaps, cohorts), saved dashboards, remote configs, experiments and their results, deeplinks, and community sentiment.
Change configurations
Edit values in remote configs, create or delete configs, choose the primary one, and make keys private or public. It can preview each change's before and after first.
Run experiments
Create an A/B test on any configuration value, ramp its rollout, read lift with a 95% confidence interval and a p-value, and end it.
Build dashboards and cohorts
Save the charts it runs, arrange them on dashboards, and save, describe, and edit player cohorts.
Instrument your app or game
Implement a tracking plan in your codebase with your coding agent, then verify each event as it arrives.
Answer from the docs
Search and read the Gamebeast docs and SDK reference, so it gets an SDK method or endpoint right the first time.
Connect in a minute
One hosted URL for every project. Sign in with your Gamebeast account; there's nothing to install.
- 1
Add the server URL to your MCP client. In Claude, go to Settings → Connectors → Add custom connector. In Cursor, add it to ~/.cursor/mcp.json.
- 2
Your client opens a browser. Sign in with your Gamebeast account; there's no token to copy and no header to set.
- 3
Choose what the connection can reach (all your projects, or only some) and what it can do (every permission your role allows, or Read only).
- Full MCP setup guide
{
"mcpServers": {
"gamebeast": {
"url": "https://api.gamebeast.gg/mcp"
}
}
}claude mcp add --transport http gamebeast https://api.gamebeast.gg/mcpcodex mcp add gamebeast --url https://api.gamebeast.gg/mcpThe same playbooks as the dashboard
The assistant's slash commands are MCP prompts, so you can start them from your coding agent.
/instrument/diagnose/weekly-review/decide-experiments/economy/liveops/segment/player-voice/dashboards/fix-instrumentation
