lss233/kirara-ai

Kirara AI is a customizable, multi-modal AI chatbot framework written in Python that lets you connect major AI language models to popular Chinese and international messaging platforms. The problem it solves is that setting up an AI assistant that actually works inside WeChat, QQ, Telegram, or Discord, complete with image generation, voice replies, and custom personas, requires connecting many different services, and Kirara AI packages all of that into one configurable system. On the AI model side, it supports OpenAI, DeepSeek, Claude, Gemini, Qwen, Mistral, Minimax, Kimi, and other major models. For image generation it supports Stable Diffusion, Flux, and Midjourney. For messaging platforms it supports Telegram, QQ bots, WeChat (both official accounts and enterprise WeChat), and Discord. It includes a web management dashboard where you can manage models, configure custom workflows (called work flows), install plugins from a marketplace, and set up conditional triggers and administrator commands. Personas (predefined character styles for the bot) can be loaded via a preset system. You would use this if you are building a Chinese or multilingual AI chatbot assistant that needs to operate across multiple chat platforms simultaneously, support voice and image capabilities, and be configured visually without editing code for every change. It is a Python project deployable via Docker.