Home
Creators
Search Recent Missing Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

SimFem Suite: v1.03 (Patreon)

Published:
2026-10-08 22:22:28
Edited:
2026-10-12 18:33:12
Imported:
Tags:

Downloads

Content

Greetings, one and all!

The major focus for v1.03 is adding in accessibility features, particularly for the visually impaired. So, give it a whirl and let me know what you think. If any issues arise or you have suggestions for improving the accessibility, let me know in the comments below.

Enjoy!

Max (Novus Operandi)

*If you enjoy this free app, you can support my efforts through buying deluxe companions or subscribing to my Patreon. Thank you!Buy Deluxe CompanionsBecome an 'Insider' Subscriber for Discounts

***

Hotfix: v1.03b

Fixed bug that could cause requested action to fail if the last chat log displayed on screen is deleted by the requested action (such as ‘Delete This Log’ or ‘Get New Response’).

Fixed bug that could cause behind-the-scenes exception if all chat logs are deleted from the screen and user then presses up/down arrow keys to “navigate” the empty field.

Minor tweaks to AI model content generation settings. In particular, made slight tweaks to reduce chaotic or nonsensical output, and increased model penalties in place to reduce repetitive output.

Changelog: v1.03

Accessibility Overhaul

Updated all major interface elements to function with TTS. Every significant interface element in the application now functions with text-to-speech compatibility in mind. This TTS functionality is optimized for 3rd party accessibility software, such as NVDA, JAWS, etc. It is not optimized for Windows Narrator. Though if Windows Narrator is detected as running, it will attempt TTS output as well. (It will not, however, actually use Windows Narrator, as full compatibility could not be achieved. Instead, it will spool up the older Microsoft Speech API and output TTS using the Microsoft Zira voice.)

Chat Log traversal with TTS output. When the chat log window or the user entry field has interface focus, you can use the arrow keys (up or down) to navigate through the chat log entries and hear each one read aloud.

Keyboard shortcuts added for major buttons. Many of the major buttons in the application have been updated to be accessible via keyboard shortcuts. Each one is a combination of the Alt key + a corresponding letter. When available, these shortcuts are visibly communicated to users via an underlined letter in the button’s label (i.e. ‘Select Character’), and audibly communicated when TTS is enabled as such, “Select Character (Alt+S)”.

Keyboard shortcuts added to launch right-click context menu. The right-click context menu, which is available when editing text, can now be launched with the following keyboard shortcuts: Shift+F10, and the Menu key.

Expanded Right-Click Context Menu w/ New Editing Actions. The right-click context menu, used for editing text, has been expanded with the following new commands: Delete, and Select All.

Easier Selection & Copying of Support Email Address. Under ‘About’ in the main interface, the support email is displayed. You can now right-click this email address and choose ‘Select All’ or ‘Copy’ to more easily paste this value into your email application.

Misc. Updates

Added term definition for “lingam massage” to app’s internal database. Complete with examples of expected behavior and actions. Now, characters should have a clearer understanding of what a lingam massage is and how to conduct one.

Increased safeguards in place to help prevent looping errors. Rarely, when asking the AI to generate creative content (such as a narrative scenario outline), it may get stuck outputting the same section of text in a loop. The safeguards in place to help prevent this behavior have now been further strengthened and optimized.

***

General Instructions

SimFem Suite: AI-Powered Romantic Companions

Bring your fantasies to life like never before, with a whole new dimension in AI-powered romance.

*CAUTION: This product contains mature content and is intended for adults*

Installation Instructions:

Download the compressed ZIP file to your desktop.

Unpack the compressed ZIP file, then run the installer found inside.

If you receive an “unknown app” or “unknown publisher” warning from Windows, do the following to get past it: Click “more info...” near the top of the pop-up window. This should reveal an option to “Run Anyway”. Click that button, and the app should run as normal. This message can appear sometimes if the publisher of the software is not a large, well-known company.

System Requirements

For SimFem:

Intel i5 14600 processor @ 3.5 GHz or equivalent

8GB system RAM

Dedicated graphics card with 12GB of VRAM

16GB available storage (6GB for app & up to 10GB for one-time AI model downloads)

Internet connection. (For initial set-up only, to download AI models for offline use)

For local Ollama framework (the engine that runs AI):

Windows 10 22H2 or newer, Home or Pro

NVIDIA 551.61 or newer Drivers if you have an NVIDIA card

AMD ROCm v7 / HIP7-capable driver stack for ROCm acceleration, or a Vulkan-capable AMD Radeon driver for Vulkan acceleration

Files

Comments

Abby

Oh, I'm so excited!! Thank you for this!

Charlie Hart

Hi Max, Thanks for the new update. I spent a few hours testing version 1.02. At first, I had a bit of trouble figuring out how everything works, but once I got the hang of it, I was able to create my own AI companion, build a custom scenario, and role-play situations I had been hoping to explore. The main thing I’m still unsure about is how the memory and internal logic work. During longer role-play sessions, the AI sometimes seems to lose track of details that were established earlier. It can mix up facts, events, or parts of the story that had already been introduced and clearly defined during previous scenes. That said, I’m really enjoying the direction of the project and I’m looking forward to seeing how it continues to develop.

Max Ingram - Author & Gamedev

Thanks for the feedback, Charlie. This is an issue I'm aware of in the program (and with the technology in general), and it's something I definitely strive to improve. It basically boils down to a balancing act between performance and quality of output. The relative "IQ" of the chat model is also a factor, and the higher the IQ, the more of a demand is placed on system resources. So the chat model used in this offline application may not always be as "sharp" as some cloud-based models you may have encountered (which are run on very powerful cloud-based hardware). The character's memories and situational awareness come from 3 sources: - The character's profile (all fields) (which is fed back into their awareness on every chat turn). - Their short_term_history log (presently 12 chat turns), which is the raw text of everything you and they have said during those chat turns (and which is fed back into their awareness on every chat turn). - Their long-term memories stored in a local database (which is dynamically queried on each chat turn using the text of your reply as a search term). If a conditional fact (such as 'the character is nude') is stated during chat turn #1, then not explicitly mentioned again, it may be forgotten on chat turn #13, because that fact has fallen outside the scope of their short_term_history logs. And if the AI character doesn't know something, their default behavior is to make something up (i.e. commonly called AI "hallucinating"). So, they may suddenly say they're wearing a dress. One trick to avoid this, which seems to work pretty well, is to periodically mention situational details in your own replies, (e.g. "I caress her nude body on the couch" instead of "I caress her body", reinserting the details of "nude" and "on the couch" in her short_term_history). The AI picks and chooses what it deems important enough to store in its long-term memories database (much like people do), so small details like clothing states are typically only stored in their short_term_history logs. There's also a phenomenon known as prompt overload. Because all of this information is being fed back into the chat model on every chat turn, sometimes things get "lost" in the shuffle. Think of it like this: If someone asks you to remember 5 things, it's easier than if someone asks you to remember 50 things. And for the AI, it's similar. They do their best, but they really do function like people sometimes, and they don't always retain everything perfectly. If you catch them in a mistake, you can try Chat Edit>Get New Response, and sometimes they'll get it right on a secondary try. Or you can simply use Chat Edit to correct their mistake yourself, save it, and then they'll be less likely to make that mistake in the future.

Charlie Hart

Thanks for such a detailed response, Max! I really appreciate you taking the time to explain how the memory system works. I'll definitely try your suggestions and see how well I can maintain a consistent, ongoing storyline. I actually think your project has a lot more potential than its seemingly simple concept might suggest. In a way, it's like an interactive erotic novel where the reader creates the story as it unfolds. But I can also see it becoming something more: a way to tell meaningful, evolving stories about two characters and their relationship, with the spicy elements being a wonderful bonus rather than the main attraction. Keep up the great work! I'm excited to see where you take this project.

Max Ingram - Author & Gamedev

Absolutely, Charlie. To me, it's like having your own personal Holodeck, only in a text-based form. I'm absolutely in love with the technology.

Daikhan

Hi. Max. I wanted to say that I really enjoyed this one. I wanted to ask how do you generate images, what do you use for it? I really liked the art style of the portraits you use. Also wanted to ask if it is possible to add image generation to SimFem. I suppose it would be a piece of work to add such feature. Any plans for that?

Max Ingram - Author & Gamedev

Hey, Daikhan. For image generation, I'm currently using Stable Diffusion XL>MJ Style XL through a cloud-based service that I subscribe to. For the specific style of output, I've been experimenting with tags such as "graphic realism art style" and "realistic comic art style". Adding image generation to SimFem Suite would be a cool feature. However, there are two major hurdles. First, technical: It would necessitate looping in a secondary AI engine to run the image generation, because Ollama (the one running chat interactions) doesn't do images. Second, ToS limitations: Patreon's current Terms of Service doesn't allow hosting applications that give users the ability to produce their own AI imagery. (I'm not sure why.) Thanks for playing. Have fun!

Daikhan

Nice to know, thanks. I hope they will change that rule at some point

Space Candy

I have two questions - as someone who's followed your work, I want to know if you plan on creating a mobile application with SimFem? Also, since I haven't used this yet, is there a way to use cloud-based models or keys? I don't know the proper term. And if there is not, do you have plans to implement this ability in the future?

Max Ingram - Author & Gamedev

No current plans for a mobile application. And everything in SimFem Suite is designed with offline sandboxing in mind. Outside of the initial model downloads, which require an internet connection, the whole app is local, offline, and self-sufficient. All data is processed though bundled engine binaries and stored on the user's own computer for increased data privacy. It's also built with a streamlined experience in mind, so you don't need to choose an AI engine or AI model. It's all handled under-the-hood. All you do is install it and start typing.

legendppk

How about you give us a set of pre-generated images instead since AI image generation isn't allowed? Exclusive download for patrons only. One for each variable like: young/mature, blonde/ginger/etc., big/small breasts, tanned/fair skin, etc. I know this feels like I'm asking for a lot but the images really do help with immersion.

Max Ingram - Author & Gamedev

Polished final images can involve a fair amount of work. For the character portraits, I'll often go through dozens of generations before I land on *just the right one*, and even then, I sometimes need to do manual edits to correct for minor issues, or even combine elements from 2 or 3 variant images to arrive at a final portrait. So, in all likelihood, I'll stick to doing those for the commercial character releases. However, I will keep your suggestion in mind.