July 2025 - Release notes
EstimAI complete rework - Intelligent Image generation - Lovense Support - Online TTS
This is a big update that I hope will be the first foundation for an even more immersive AI experience.
A huge thank you to all of you supporting the development of Eudaimonia! I'm so grateful!
I also want to mention that the awesome intelligent image generation feature in this update is mostly the work of Shannon. I think it takes the immersion a solid step forward!
As always, feel free to reach out (contact@daimonia.app) if you run into any weird behavior, bugs, or just want to share feedback.
EstimAI Overhaul
I've made a lot of structural changes to the whole scenario/persona system in EstimAI and worked on many improvements in general.
The main idea behind this is that I believe there's huge potential to create rich and immersive experiences with AI. We're reaching a point where models are so good that it's possible to craft elaborate, complex, and rich scenarios.
Three key points:
Scenarios and personas can now be published and shared. Plus, there's a new Scenario Editor!
I can't wait to see all the creative and amazing ideas from the community! I've added a full editor to give creators a practical tool.
Now you can publish your scenarios to share your best creations with everyone. You can stay anonymous if you want, or set an ‘author name' in your account settings.
Any content published will go through automatic moderation, then human verification if it gets flagged. The moderation is mostly to prevent illegal content from being shared publicly.
Trigger / Event system for your scenarios
A big change in this update is the new Trigger - Event system.
You can now create "triggers" that, when their conditions are met, trigger "events" during a scenario. For example, you can now set up distinct phases within a single scenario, with their own specifications and transitions, without overloading the prompt.
I've updated The Succubus and Interrogatoire as examples of how to use it.
Here is guide to show how it works, since I think it might be a bit confusing at first.
Below is a short demo of a simple phase change example combined with the new 'intelligent generation' feature.
Phase 1:The scenario suddenly shifts from: {{char}} is a submissive woman that will do anything you ask. To Phase 2:The atmosphere shifts, darkness grows, the room is bathed in ominous red lighting. {{char}} suddenly reveals her true dominant nature and turns into a succubus.
I believe this has a lot of potential to create a more complex and creative experience with AI, while reducing its inherent flaws like repetitiveness, not driving the plot forward, and relying on the same clichéd sequences.
This update is just the foundation, I've added a few primary tools, but I intend to add more to the editor.
I would also like to create a large "Dungeon" scenario, in which you would progress through different rooms, each offering a special encounter with a different setting.
Easier usage
EstimAI looks like a complicated mess, probably not very appealing. I wanted to make the setup easier and more attractive. The idea is that now you just pick a scenario with a fitting persona, select a model, click 'Apply Recommended Settings', choose an appearance, and play.
TLDR Features
- New scenario Editor
- Scenario and persona can now be published.
- A new complex system to improve scenario quality has been added: Trigger → Event.
- Scenario and persona are now two distinct entities, meaning you can pick any persona for any scenario.
- A recommended persona feature still lets you link a specific persona to a specific scenario.
- Two revisited scenarios: Succubus & Interrogatoire, now with examples of how to use Triggers & Events.
- EstimAI and Appearances (especially for image generation) are now more deeply linked. You can auto-apply the description + outfit of the chosen appearance to the prompt.
If you use this setting, remove any {{char}} descriptions in your scenarios/personas. - Scenarios now have a location, automatically added to the prompt. You can pick between the character's background or the scenario's location in the Appearance Form for image generation.
Intelligent Image generation (EstimAI - For subscribers)
Shannon designed, experimented, and built a new system to sync images with answers, bringing a lot more consistency and fidelity.
I integrated it into Eudaimonia and optimized it. I believe this is a big step forward for immersion, and it works really well with the new scenario system.
There are still some flaws since it's highly dependent on both AI (text + image generation), but I've been really impressed by it.
I added a module to the game that allows you to edit the prompt image, so you can remove unnecessary details, keep the model on track, etc. You can open it by clicking on the icon .
Anime/illustration is highly recommended over realism for this feature to work properly.
The video above shows a short demo of this feature in action.
This feature is only available to subscribers, as it sends multiple complex requests and requires a good balance between intelligence and speed, making it unusable a local model.
Lovense Sex toys support
Another big feature of this update is the creation of a second portal called 'Sextoys' that lets you play with connected Lovense sex toys.
I plan to add support for more toys from different brands soon (probably using the Intiface/buttplug.io API).
And in the more distant future, I would like users to be able to mix any kind of sextoys and e-stim devices to create their perfect setup.
Online TTS (for subscribers)
I've added support for Online TTS for subscribers. If you can't or don't want to use the local TTS feature, this one is much easier to use.
More info about rates and credits here.
This API is expensive, which is why the available credits are so limited. Still, I find that the quality and customization options are better with the local TTS feature. I'd recommend using this feature if you're looking for simplicity or if you can't use the local TTS feature.
EstimAI Improvements
A lot of improvements have been made to EstimAI, especially two new features: Online Utilities & Debug Mode.
Online utilities: Lets subscribers use their favorite local model along with all the utility features (like intelligent image generation or nudity detection) with the Online API.
Debug mode: The scenario won't be played, as sequence of lines being displayed, the answer will be generated and then displayed directly as text (like a more classic text chat interface). This allows you to debug your scenarios more easily. There's also a new icon you can click to follow the entire scenario.
Stims: The stim system has been reworked to provide more variety, making it more meaningful and consistent.
The new system was built with modularity in mind, as I want to let scenario creators provide custom sets of stims for their scenarios later on. I'll probably also add a way to filter playable stims in scenarios (using the trigger/event feature).
New models: Since scenarios are getting more complex and detailed, I've added support for larger models like Claude 4 Sonnet and Gemini 2.5 Pro. There are also constantly awesome new models (like Kimi-k2, Qwen3, or Deepseek) that deserve a spot.
Please check the prices of the models beforehand, as some are quite expensive and can quickly exhaust your credits (Claude 4, Gemini 2.5 Pro).
- Online utilities
- Debug module
- New stim system
- New settings (OOC message, max shocks, More intrinsically linked to appearances and image generation)
- New presets
- Better sentance segmentation
- New models
Image generation Improvements && Ilustrious Support
A few improvements have been made to image generation.
Illustrious support has been added (both anime and realism). Illustrious has been highly regarded in the Stable Diffusion community, and it was about time I finally added support. You can find a small list of interesting Illustrious checkpoint merges below:
List of recommanded models for ILLUSTRIOUS (anime):
List of recommanded models for ILLUSTRIOUS (realistic):
- Illustrious Support
- Type of generation + face detailer option have been moved to the general image generation settings
- Biglove bug fixes
Miscs
- Infrastructure: The community is growing, so I've made a lot of important changes to the general infrastructure. Most of them are invisible, but they should ensure better performance and a more stable app.
- A few Coyote3 improvements (more settings, less brutal stim changes, smoother calibration).
- Many small UX changes to make things a bit easier and clearer.
- And plenty of other small things that I don't remember.