Skip to content

Python Manta

Python bindings for the dotabuff/manta Dota 2 replay parser

AI Summary 
Python Manta lets you parse Dota 2 replay files (.dem) in Python. Install with `pip install python-manta`. The library wraps the Go-based manta parser and provides:
  • Single-pass parsing for header, game info, and messages
  • 272 message callbacks for any replay data
  • Game events (364 types), combat log, modifiers, entity queries
  • Pydantic models for type-safe data access

Quick start: from python_manta import Parser; parser = Parser("match.dem"); result = parser.parse(header=True)

What is Python Manta?

Python Manta provides Python access to the Manta Go library for parsing Dota 2 replay files. All parsing is done by Manta - this library provides:

  1. CGO bindings - Wraps the Go parser as a shared library
  2. Pythonic API - Clean, intuitive interface via Parser
  3. Type safety - Pydantic models for all parsed data
  4. Comprehensive access - All 272 Manta callbacks + specialized APIs

Installation

pip install python-manta

Pre-built wheels available for Linux, macOS (Intel + Apple Silicon), and Windows.

Version Pinning

Always use the latest release for your target Manta version:

# Latest release for Manta 1.4.5.x (recommended)
pip install "python-manta>=1.4.5,<1.4.6"

# Or use compatible release operator
pip install "python-manta~=1.4.5"

See Getting Started for version format details.

Quick Example

from python_manta import Parser

# Create parser for a specific demo file
parser = Parser("match.dem")

# Parse header and game info in a single pass
result = parser.parse(header=True, game_info=True)

# Access header data
print(f"Map: {result.header.map_name}, Build: {result.header.build_num}")

# Access draft picks and bans
for pb in result.game_info.picks_bans:
    action = "PICK" if pb.is_pick else "BAN"
    print(f"{action}: Hero {pb.hero_id}")

# Parse messages separately
result = parser.parse(messages={"filter": "CDOTAUserMsg_ChatMessage", "max_messages": 100})
for msg in result.messages.messages:
    print(f"[{msg.tick}] {msg.data}")

Features Overview

Feature Collector Description
Header header=True Match metadata (map, build, server)
Game Info game_info=True Draft, players, teams, league
Messages messages={...} Any of 272 message types

Documentation