Hey everyone,
I’ve been developing a passion project called StrokeGPT, a free, open-source Windows app that transforms your Handy into an adaptive, intelligent, and fully generative erotic partner.
My goal is to create an experience as close as possible to giving a real, thinking partner control of your Handy, running a local language model on your machine.
What’s New in Version 1.2.5?
This is a major update focused on user experience and new features.
Complete UI Overhaul: The interface has been redesigned into a modern, two-column layout. The chat window is now much larger, with all setup controls moved to a dedicated sidebar for a cleaner, more ergonomic experience.
ElevenLabs Audio Integration: The app now fully supports ElevenLabs text-to-speech. A new "Audio Settings" panel allows you to use your API key to load your personal voice library for a fully voiced session.
Advanced In-Session Memory: The AI's short-term memory has been massively upgraded. It now automatically summarizes the conversation as it goes, allowing it to recall context from the beginning of a long session without amnesia or slowing down your computer.
Project Now on GitHub: StrokeGPT has a new home! The project is now hosted on GitHub, which means no more fussing with different download links. You can always find the latest version, see what's changed, and track development in one place.
Upgraded AI Model: The recommended model has been updated from mistral-openorca to llama3:8b-instruct-q4_K_M. It's significantly smarter and faster at following instructions.
Depth Understanding: The AI now understands max depth and overall depth on a much deeper level.
Download & Installation
The .exe installer no longer exists. StrokeGPT is now distributed as source code for a more transparent and reliable setup.
Get the code from the new GitHub page:
Click the green < > Code button, then “Download ZIP”.
Installation Guide:
You will need to install Python and the AI server, Ollama. This is a one-time setup.
Install Python:
Go to python.org/downloads and install the latest version.
IMPORTANT: During installation, check the box that says "Add Python to PATH".
Install Ollama (The AI Server):
Go to ollama.com and install the application. Make sure it's running before you proceed.
Set Up the Project:
Unzip the files you downloaded from GitHub into a new folder.
Open a terminal in that folder (type cmd in the folder's address bar and press Enter).
In the terminal, run the following commands one by one:
Install required libraries:
pip install -r requirements.txt
Download the new AI model:
ollama pull llama3:8b-instruct-q4_K_M
(This will take several minutes and requires a good internet connection).
Run StrokeGPT:
In the same terminal, run:
Bash
python app.py
The server will start. Open your web browser and go to the address it shows: http://127.0.0.1:5000
Key Features & Troubleshooting
Dynamic Persona & Mood Engine: Describe any partner and the AI adapts its personality. It reacts to the subtext of your messages, creating a dynamic and realistic emotional arc.
Generative "Dancer" & "Milking" Modes: All automated stroking is created on the fly. The AI choreographs a unique dance based on its mood and can be instructed to enter a relentless, climax-focused milking mode.
Self-Learning: Use the "Like" button to teach the AI what you enjoy. It will learn and save your favorite moves for future sessions.
AI is Slow or Times Out? This usually means the AI is running on your CPU instead of your GPU. The simplest fix is to open app.py, find the line with timeout=60, and change it to timeout=180.
I’d love to hear your feedback. Thanks for your support!
If you’d like to support the project, you can do so here (donations go toward a Kiiroo Keon for future development): Kofi.
