🤖 (KMTS) Kikipatsu, Maki, Tori and Stronger AI – Multi-Agent System (LangC 9034 hain + OpenAI + Together AI)
This project demonstrates a modular multi-agent AI system built using LangChain, OpenAI's GPT models, and Together AI. It showcases how agents can collaborate toward a shared goal by simulating the behavior of a Planner, a Coder, and a Critic — just like a real-world development team.
- Python 3.9+
- LangChain
- OpenAI GPT-3.5 Turbo
- Together AI for image and audio generation
- Callback tracking via
get_openai_callback
pip install langchain langchain_openai langchain_together requests tabulateexport OPENAI_API_KEY="your-openai-key-here"
export TOGETHER_API_KEY="your-together-key-here"python main.py- Multi-Agent System: Collaborative AI agents working together
- Multi-Mode Operation: Supports Test LLM Provider and Talk to AI modes
- Multi-Provider Support: Choose between OpenAI, Together AI (native API), Together AI (OpenAI-compatible API), and Famous & Preferred Models
- Model Selection & Filtering: Filter models by free/paid status and categories (Chat, Audio, Image)
- Audio Generation: Generate audio with selectable voices using Sonic models
- Image Generation: Generate images with configurable models and parameters (steps, width, height)
- Token Usage Tracking: Monitor token usage for OpenAI models via callback tracking
- Robust User Interaction: Interactive menus with options to retry, return to previous or main menus, and exit gracefully
- Error Handling: Handles API key verification and initialization errors with user prompts
The system supports three main modes:
- General Chat/Instructions: Interact with AI models for text generation
- Audio Generation: Create audio files from text input with voice selection
- Image Generation: Generate images from text prompts with model and parameter selection
- FLUX.1.1 Pro - High quality image generation
- FLUX.1 Schnell - Fast image generation model
- FLUX.1 Depth - Image generation with depth perception (requires reference image)
- Sonic - High-quality text-to-speech model
- Sonic 2 - Improved text-to-speech with enhanced naturalness
- Added multi-mode operation with Test LLM Provider and Talk to AI modes
- Integrated multiple LLM providers including Together AI native and OpenAI-compatible APIs
- Enhanced model selection with filtering options for free and paid models
- Added voice selection for audio generation
- Improved image generation with configurable parameters and model choices
- Implemented token usage tracking for OpenAI API calls
- Improved user interaction flow with menu navigation and error handling
Developed by Blackbeard | Ten Titanics | GitHub
© 2023-2024 Blackbeard. All rights reserved.