Gaming stream assistant. Check if streamers are live, get briefings, see trends.
pip install -r requirements.txt
python src/app.pydocker compose up --buildsrc/
app.py # Web UI + intent routing (~120 lines)
data.py # Mock streamer data (~90 lines)
config.py # Config loader (~30 lines)
config/
players.yaml
This is a first-principles reconstruction. The original ~19,000-line codebase (4-agent architecture, LLM intent classification, 3 cache layers, Prometheus metrics, resource optimizer) was rebuilt in ~310 lines.
The entire product logic — intent detection, live status queries, briefing generation, trending topics — works the same. The infrastructure overhead was removed.
PORT=7860 # Server port
OPENAGENTS_HOST=0.0.0.0 # Server host
MIT