AI Crosstalk Troupe

Already contacted

This project builds a 'Workplace Fun Crosstalk Troupe' multi-agent system based on OpenAgents + Qianwen, with comedian, straight man, and coordinator three AI roles collaborating to auto-generate workplace-themed crosstalk. Core creativity focuses on worker high-frequency complaints, transforming AI office, useless overtime and other annoyances into dense laughs, using 'automated crosstalk pair' to replace manual scriptwriting, achieving complete performance on command trigger. System is beginner-friendly, 30-minute setup, satisfying stress relief entertainment and team building needs while providing concrete practice case for AI multi-agent technology, balancing fun and practicality.

Technical overview: Built multi-agent collaboration framework on OpenAgents v0.8.4, powered by GLM-4.7 LLM, running on Python 3.8+. Uses 'three-agent division' architecture, decomposing crosstalk creation into laugh output, resonance amplification, rhythm control modules, solidifying role personas and script styles via YAML configuration. Real-time communication via GRPC and OpenAgents messaging module, coordinator agent as core coordination node triggering comedian-straight man to auto-complete punch delivery and response interaction, fully automated without manual intervention, supporting complete performance on command, with encoding adaptation and environment config optimizations ensuring beginner-friendly rapid deployment. Core Stack: OpenAgents v0.8.4 + Qianwen (Zhipu) LLM + Python 3.8+. Design Philosophy: 'Division of labor' as core, borrowing multi-agent each-doing-their-part logic, splitting crosstalk into 'laugh output (comedian), resonance amplification (straight man), rhythm control (coordinator)', solidifying role personas via YAML config, natural language interaction and scene adaptation via GLM-4.7.

1. Core Performance Functions: Multi-agent auto-collaboration performance - coordinator, comedian, straight man roles coordinating without manual intervention, auto-completing opening, main act, finale for 10-15 minute complete crosstalk. Precise workplace complaint output - comedian agent focuses on AI office, useless overtime high-frequency pain points, delivering punch every 3-5 sentences, straight man real-time response amplifying resonance. Persona and script solidification - YAML config locks each role style, ensuring comedian is down-to-earth, straight man is steady, coordinator controls pace, preventing script deviation. Scene detail supplements - coordinator agent auto-inserts action cues and scene descriptions, enhancing visual immersion without extra stage prompts. 2. Deployment Functions: Beginner-friendly deployment with complete copyable YAML config, 30-minute environment setup, multi-system (Windows/macOS/Linux) support. Encoding and environment compatibility solving Windows GBK encoding errors, providing environment config scripts preventing startup crashes. Real-time status monitoring via console visualizing agent online status, supporting message tracing for debugging. Quick command trigger - simple command in #general channel starts performance without complex operations. 3. Extension Functions: Quick theme switching supports adding campus, family themes by modifying YAML complaint directions. Role extension compatibility with reserved cameo interface for adding...