OpenDota MCP Server
# OpenDota MCP Server
[](https://smithery.ai/server/@asusevski/opendota-mcp-server)
A Model Context Protocol (MCP) server implementation for accessing OpenDota API data. This server enables LLMs and AI assistants to retrieve real-time Dota 2 statistics, match data, player information, and more through a standard interface.
## Features
- Access player profiles, statistics, and match history
- Retrieve detailed match information
- Look up professional players and teams
- Get hero statistics and rankings
- Search for players by name
- And more!
## Installation
### Installing via Smithery
To install OpenDota API Server for Claude Desktop automatically via [Smithery](https://smithery.ai/server/@asusevski/opendota-mcp-server):
```bash
npx -y @smithery/cli install @asusevski/opendota-mcp-server --client claude
```
### Manual Installation
```bash
# Clone the repository
git clone https://github.com/asusevski/opendota-mcp-server.git
cd opendota-mcp-server
# Option 1: Automated setup (works with bash, zsh, and other shells)
./scripts/setup_env.sh
# Option 2: Manual installation with uv
uv add pyproject.toml
# For development dependencies
uv pip install -e ".[dev]"
```
## Usage
### Setting up your environment
1. (Optional but recommended) Create an OpenDota API key at https://www.opendota.com/api-keys
2. Set your API key as an environment variable:
```bash
export OPENDOTA_API_KEY=your_api_key_here
```
### Running the server directly
```bash
python -m src.opendota_server.server
```
### Running the server with Claude Desktop
Follow this: https://modelcontextprotocol.io/quickstart/user
If you use WSL, assuming you have cloned the repo and set up the python environment, this is how I wrote the claude_desktop_config.json:
```json
{
"mcpServers": {
"opendota": {
"command": "wsl.exe",
"args": [
"--",
"bash",
"-c",
"cd ~/opendota-mcp-server && source .venv/bin/activate && python src/opendota_server/server.py"
]
}
}
}
```
### Using the example client
```bash
python -m src.client
```
### Specific tools included:
- get_player_by_id - Retrieve player information by account ID
- get_player_recent_matches - Get recent matches for a player
- get_match_data - Get detailed data for a specific match
- get_player_win_loss - Get win/loss statistics for a player
- get_player_heroes - Get a player's most played heroes
- get_hero_stats - Get statistics for heroes
- search_player - Search for players by name
- get_pro_players - Get list of professional players
- get_pro_matches - Get recent professional matches
- get_player_peers - Get players who have played with a specified player
- get_heroes - Get list of all Dota 2 heroes
- get_player_totals - Get player's overall stats totals
- get_player_rankings - Get player hero rankings
- get_player_wordcloud - Get most common words used by player in chat
- get_team_info - Get information about a team
- get_public_matches - Get recent public matches
- get_match_heroes - Get heroes played in a specific match
## Security
[](https://mseep.ai/app/asusevski-opendota-mcp-server)
## License
MIT
TDQS
Scored across 17 tools
Most tools have distinct purposes targeting different resources (heroes, matches, players, teams, pro scene). However, some player-focused tools like get_player_heroes and get_player_rankings could potentially overlap in retrieving hero-related player data, though descriptions clarify their specific focuses.
All tools follow a consistent verb_noun pattern with get_ prefix and snake_case throughout. The naming is highly predictable and readable, making it easy to understand each tool's function at a glance.
17 tools is slightly high but reasonable for a comprehensive Dota 2 statistics server. The count supports detailed player analysis, match data, hero stats, and professional scene coverage without being overwhelming for the domain scope.
The toolset provides excellent coverage of the OpenDota API domain with comprehensive player analysis (stats, matches, heroes, peers, rankings), match data (public, pro, detailed), hero statistics, team info, and search functionality. No obvious gaps exist for typical agent workflows.