SYSTEM OVERRIDE ACTIVE

OPENRA-RL

OpenEnv environment for training AI agents to play Red Alert through the OpenRA engine. Connect via WebSocket or HTTP, send actions, observe the battlefield.

WATCH AI PLAY DOCUMENTATION LEADERBOARD

Endpoints

API DOCS

Interactive Swagger UI with all REST and WebSocket endpoints.

/docs →

HEALTH CHECK

Server status and readiness probe for monitoring.

/health →

ENV SCHEMA

JSON schemas for actions, observations, and game state.

/schema →

Connect to Environment

Use the Python client to connect, reset the environment, and step through the game loop. Works with both local Docker and this HuggingFace-hosted server.

API REFERENCE

terminal

$ pip install openra-rl

from openra_env.client import OpenRAEnv
from openra_env.models import OpenRAAction

url = "https://openra-rl-openra-rl.hf.space"

async with OpenRAEnv(url) as env:
    obs = await env.reset()
    while not obs.done:
        action = your_agent(obs)
        obs = await env.step(action)