When i try this it says “C:\Compilers\mingw64\bin\python.exe: No module named pip” 
I’m honestly not sure, sorry. It’s not something I’ve come across yet. If I find a solution, I’ll tag you with it here. 
I’ve added an FAQ to the OP which should help answer common questions.
If I’m ‘chatting’ does it make sense it doesn’t update the strokes after one chat from me? For example, if I say “do x with variations”, it just does “x” once, and then waits for further input from my side instead of changing it up every so foten
Yes. The model can currently only send one set of vectors to the Handy servers then to the handy/one per message. Automodes are the same, too. The model can’t understand any requests asking for variations as it’s understanding boils down to,
“The user wants a looping up/down at this speed at this part of their cock.”
This means you have a python installation that did not bundle pip.
I don’t know exactly why but it can happen. These docs have a few solutions for your situation, you will have to try them to figure out what the issue is.
Thank you. I appreciate your help. 
For those curious about the work I mentioned a few days ago about splitting the app file into modules. It’s making great progress. Obviously one gigantic app file is okay on paper but it’s an absolute bitch to keep track off and the more I cram in the more fucky it gets, no matter how well it’s noted. This is partly down to me being a messy amateur who learned his shit way too many moons ago. But, this is what we’re looking at for the future.
From your perspective nothing will change. You’ll just run the app file as normal, but editing different aspects of the app will now be much, much easier for me and anyone else interested in having a fiddle 
.
The UI and current features have also undergone quite the overhaul. Here’s what exactly:
- The AI can now have a profile picture (from your pc) that’ll appear alongside its messages.
- Everything has been smoothed.
- The sidebar can now collapse (not featured in this pic).
- The sidebar can now scroll.
- There’s now a typing indicator for the AI messages.
- There’s now more natural message animations.
- You can now name the bot and have it’s name show up with messages (not featured in this pic).
- You can now message the AI during auto/milking/edging mode and it’ll not only take your words into consideration, but also continue the current mode, too.
- Edging Mode has been greatly expanded. I don’t want to spoil this so you’ll see.
- Something to do with old codes and new test subjects being welcome?
I would love to work with an image generation service that’d be willing to work with me and this project but I’ve only been turned down so far. Generating locally is a no-go but working with an existing service (likely for a price) would work well and be very easy to implement.
2 Likes
Thanks for your time, but it seems too complicated for me.
1 Like
No problem. I get that fully.
I did, but I did it again and now it works. I think because I only updated Python after it didn’t work and didn’t run that command again afterwards…? IDK, everything from Github is witchcraft to me.
The app is now modular, baby! Which makes changing/adding stuff much easier!
Despite all the internal upgrades and new features, the way you run the application has not changed at all. You still start the server from your terminal with the exact same command as before:
python app.py .
=== v1.5 - The Refactor & Interactivity Update ===
It’s a doozy…
- Major Architectural Overhaul: The entire application was refactored from a single mahoosive file into a multi-module structure. This makes it more stable and easier to update in the future. (I wept a little.)
- New Feature: Advanced Interactive Edging Mode! The old, predictable loop is gone. Edging is now a proper game:
- A new “I’m Close” button lets you signal the AI when you’re near the edge.
- The AI is now unpredictable and will react to your signal by choosing to pull back, hold you on the edge, or even push you a little further.
- Chat messages sent during any mode (Auto, Milking, Edging) will now influence the AI’s behavior in real-time.
- New Feature: AI-Aware Edging Timer - A timer now appears during edging, and the AI is aware of it, allowing it to comment on your performance.
- New Feature: Custom AI Name & Profile Picture - You can now set a custom name and upload a profile picture for the AI in the sidebar. The name replaces “BOT” in the chat.
- UI Overhaul: The entire interface has been polished with a more modern design, better fonts, softer elements, and smoother animations.
- Collapsible Sidebar: The sidebar can now be hidden to focus on the chat.
- Added an AI typing indicator and animated messages to make the chat feel more alive.
- The title has been updated to the much sleeker “StrokeGPT”.
- Added a few secret chat commands and naming conventions for users with particular tastes in classic gaming.
- CRITICAL SAFETY FIX: The “Stop” command is now top priority and will correctly interrupt all modes, including the new interactive edging mode.
- Fixed a persistent bug that caused the onboarding wizard to appear on every startup.
- Fixed numerous issues where the AI would stop moving if it received an incomplete command from a user’s request.
- Massively improved the AI’s long-term memory. It should no longer get confused about who is who or forget details from the chat (still not perfect).
- The “Like” button now actually works and saves patterns correctly.
- Tons of other little bits and bobs.
If you’re running an older version of the app, just download the entire directory and replace everything you have with all the new files/folders. Keep your my_settings (or start a new one, I’m not your boss) file and start to app as normal!
If you’re new, just follow the instructions in the README! file. It’s really not that hard, and most of it is just copy pasting commands.
Thank you for all the support. This is a tough project because it’s become much larger than I first anticipated.
6 Likes
amazing, this could make modding a lot easier. 
but dang, should have waited with my UI tweaks
back to forking i guess 
2 Likes
Sorry, homie. But amazing job! Love the additions. Fork away!
Yeah, should be way easier now, and no more file layout changes from now on. I’m resting for a few days while I recover from re-connecting waaaaay too many servered connections from that big-assed app file lol.
I wonder if a thread for StrokeGPT “mods” would be good or even just for variations? It’s got way bigger than I expected 
1 Like
Enjoying the update man. Trying out edging mode though and pretty much all strokes are just fast and only on the tip. Not sure if it’s a settings limit mode on my end or not. Auto seems to be pretty much the same as well.
3 Likes
Thank you. I will check this out tonight. It may be that I again forgot to disable a few limits I had on while testing lol. Sorry about that. If it’s a small fix, I’ll post the update straight after.
Double checked that one line in the handy_move function and it looked fine I think, not sure if you put in a limit somewhere else. Could definitely just be my settings.
I’ll have a look. I may have also not correctly connected everything back together when I split everything (not how I normally work). Either way, I’ll have a look for you later 
.
Are you min/max depth settings okay in your settings file?
1 Like
I think so— have them set to 3/70 min/max.
First of all, thanks for what you do, I tried the app a few days ago…
I’ve noticed that the voice prompt also repeats words in parentheses (commands or other). It starts rhythmically right away…you should make it slow at first…or have it make quick or slow movements that stop immediately. At the word “stop” in the sentence it stops while chatting.