MVPandey/DTS

🌳 MCTS-inspired parallel beam search for conversation optimization. Explore multiple dialogue strategies simultaneously, stress-test against diverse user personas, score with multi-judge consensus, and discover winning conversation paths that single-shot LLMs miss.

45
/ 100
Emerging

Implements parallel beam search with multi-turn conversation rollout, forking each strategy across diverse simulated user intents (cooperative, skeptical, resistant, etc.) to stress-test robustness. Multi-judge consensus scoring with median voting prunes underperforming branches; supports both comparative (force-ranked siblings) and absolute (independent 0-10) scoring modes. Integrates with OpenRouter/OpenAI for LLM calls and includes real-time web-based visualizer for tree exploration, strategy playback, and evaluation breakdowns.

No Package No Dependents
Maintenance 10 / 25
Adoption 7 / 25
Maturity 13 / 25
Community 15 / 25

How are scores calculated?

Stars

35

Forks

6

Language

Python

License

Apache-2.0

Last pushed

Jan 18, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/agents/MVPandey/DTS"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.