MyGirlGPT enables users to create a personalized AI girlfriend running locally on their own server, offering control and privacy. It targets individuals seeking a customizable AI companion with unique personality, voice, and image generation capabilities, integrated via Telegram.

How It Works

The architecture comprises a Telegram bot for user interaction, an LLM server for conversational responses and image generation requests, a TTS server for voice output, and a text-to-image server leveraging Stable Diffusion WebUI. The LLM server acts as the central intelligence, processing messages, invoking TTS for speech, and calling Stable Diffusion for image generation based on conversation context.

Quick Start & Requirements

• Install/Run: Requires starting Stable Diffusion WebUI (with --api and optionally --listen), the LLM Server (using text-generation-webui), and the TTS Server (using Bark). Finally, start the TelegramBot, configuring server addresses.
• Prerequisites: Stable Diffusion WebUI, text-generation-webui, Bark TTS, Telegram Bot Token, Python environment. Specific LLM model (e.g., TehVenom/Pygmalion-Vicuna-1.1-7b) and TTS model (Bark) are needed.
• Resources: Requires significant compute resources for running LLM and Stable Diffusion models locally.
• Docs: RunPod Guide

Highlighted Details

• Telegram integration for seamless chat interaction.
• Local LLM (text-generation-webui) for enhanced privacy.
• Customizable personality and voice generation via Bark.
• Selfie generation using Stable Diffusion WebUI API.

Maintenance & Community

• Open to pull requests; suggests opening an issue for significant changes.
• Powered by Stable Diffusion web UI, Bark, text-generation-webui, and TehVenom/Pygmalion-Vicuna-1.1-7b.

Licensing & Compatibility

• MIT License.
• Permissive license suitable for commercial use and integration with closed-source projects.

Limitations & Caveats

The project is described as an AI agent deployable on OpenDAN-Personal-AI-OS, but the primary README focuses on a self-hosted, multi-component setup. Long-term memory and video message features are listed on the roadmap, indicating they are not yet implemented.