Skip to content

GodotH/zero-chat

Repository files navigation

zero-chat

High-Precision Multi-Agent Reasoning Interface

Zero-Chat is a web interface implementing the MAKER framework (Massively Decomposed Agentic Processes) to solve complex tasks with "zero error" reliability. It uses Google's Gemini 3 Pro model to decompose tasks into atomic steps, generate multiple candidate solutions for each step, and vote on the best outcome.

🔗 Concept Paper

This project is an implementation of the concepts discussed in:
Solving a Million-Step LLM Task with Zero Errors
Meyerson et al., Cognizant AI Lab (2025)


Key Features

  • MAKER Framework: Automatic task decomposition, candidate generation, and voting.
  • Multimodal: Analyze Images, PDFs, and Text files with rigorous verification.
  • Zero Error Protocol: Red-flagging logic detects and discards hallucinated or lazy responses.
  • Cost Transparency: Real-time tracking of token usage and estimated cost.
  • Resilience: Auto-save history, zombie-task cleanup, and error boundaries.

Disclaimer

This is a community project created for educational and experimental purposes. It is not an official product of Cognizant AI Lab or Google.

License

CC BY-NC 4.0 (Creative Commons Attribution-NonCommercial 4.0 International) Free for non-commercial use, modification, and distribution.

About

zero chat is a chatbot test implementation of MAKER system based on the Solving a Million-Step LLM Task with Zero Errors paper.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors