Skip to main content
MCP server CX Codex

Batru — Claude Code MCP server: Dota 2 and Deadlock predictions

Gives your agent calibrated win-rate predictions, pick suggestions and matchup tables for Dota 2 and Deadlock. Handy when you want draft advice instead of guesses from memory.

Install into Claude Code · Platform CLI: remote server All methods and platforms ↓
claude mcp add --transport http batru-mcp https://batru.gg/mcp
Open on GitHub

What Batru — Competitive Game Intelligence for AI Agents does

The server sits on top of batru.gg's public API and runs no model of its own: every figure comes from the same production model the site serves. It exposes 31 read-only tools covering Dota 2, Deadlock and Marvel Rivals. The core ones are predict_dota_winrate, recommend_dota_pick, get_dota_counters and predict_deadlock_draft.

Hero names go in however you like, aliases such as am or anti mage included, and get normalised on the way in. The backend drops names it does not recognise without complaining, so call lookup_hero first — skip that step and the numbers drift. Partial drafts are fine; an empty one returns 50/50.

There is nothing to install for the canonical setup: point your client at the hosted endpoint https://batru.gg/mcp. The PyPI package uvx batru-mcp is frozen at an early version with a smaller tool set, though it proxies the same live API. Running it locally needs uv and Python 3.12+, and the API base is set through BATRU_API_BASE.

What it helps with

  • score a Dota 2 draft that is only half filled in
  • shortlist three heroes for the next pick
  • check who a given hero beats and loses to
  • read the odds in a Deadlock 6v6 draft
  • map a hero alias to its canonical name

Example prompts for the agent

  • Predict the win rate for this draft: we have Anti-Mage and Invoker, they have Phantom Assassin
  • Who should we pick third against am and lina?
  • Show me the matchups for Anti-Mage, first 12

How to install Batru — Competitive Game Intelligence for AI Agents

1Platform CLI: remote server

claude mcp add --transport http batru-mcp https://batru.gg/mcp

2Platform CLI

claude mcp add --env BATRU_API_BASE=<BATRU_API_BASE> --transport stdio batru-competitive-game-intelligence-for-ai-agents -- uvx batru-mcp

Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from batru-mcp <command>

Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more

Questions about Batru — Competitive Game Intelligence for AI Agents

How do I install Batru in Claude Code?

Add the hosted endpoint: claude mcp add --transport http batru-mcp https://batru.gg/mcp. For the local PyPI route: claude mcp add --env BATRU_API_BASE=<BATRU_API_BASE> --transport stdio batru-competitive-game-intelligence-for-ai-agents -- uvx batru-mcp.

What does Batru do?

It hands your agent tools for win-rate prediction, pick suggestions and matchup tables in Dota 2 and Deadlock. All of it is read from batru.gg's live API; nothing is computed locally.

How is the hosted endpoint different from the uvx batru-mcp package?

The hosted endpoint https://batru.gg/mcp stays in sync with the site and carries the full tool set. The PyPI package is frozen at an early version with fewer tools, even though it proxies the same API.

Similar MCP servers

All MCP servers for Agent setup: memory, context, orchestration →

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor