April 2026 (April 30 is still April) - Release notes

Here is the April 2026 release! I'm happy to continue the development of Eudaimonia, bringing new features and trying to make the experience even more immersive and interactive!
I hope you'll enjoy it!

This release contains:
- The first Eudaimonia character Loras
- EstimAI "Modules" feature
- A visual overhaul
- New features for the EstimAI scenario editor
- And more!

As always, feel free to reach out (contact@daimonia.app) if you run into any weird behavior, bugs, or just want to share feedback.
I don't always have the time to implement all your great suggestions right away, but I'm not forgetting about them, most are waiting patiently in my novel entitled "Todo list".

Regarding the Intiface integration, a user from the community (who already built the local-TTS server!) has developed a bridge between Eudaimonia and The Handy's online API. If you've run into issues with the Bluetooth connection, or if you're on an Apple OS, this may be a life-saver for you!
You can clone it from GitHub: The Handy Bridge!

Eudaimonia's Character LoRAs:

You can now download LoRA models of Eudaimonia's characters: Natsuki, Stalla, and Nahama!
You can download them from Civitai, drop them into ComfyUI's loras folder, and use them easily in your scenarios. Two versions of each have been trained: Illustrious Anime and Illustrious Realistic.
I may train more versions in the future (Z-image Turbo, LTX2.3, etc.).

Stalla will be introduced with a new scenario coming next week!

You can find all characters loras here.

EstimAI "Modules"

This isn't a feature in itself, but more of a new way to use EstimAI that will allow for a lot of new possibilities and improvements in the future.
I want to give the AI more and more tools to interact with the user and the environment, and this is the first step in that direction.
A module is a feature that you can enable or disable according to your needs. In the end, you'll be able to combine multiple features and agents to create a richer experience in EstimAI.

I've developed four new modules for this release:

Quick Replies

You can now fully edit the common "Quick Replies" available in EstimAI, and you can create specific ones for your scenarios. These Quick Replies can also be changed between different phases of a scenario, using the Trigger/Event system.
An option also exists to make a scenario only use Quick Replies, which can be interesting for more interactive and dynamic scenarios where you don't want the user to have to write inputs, but just click on predefined options.

Here are the customisable options:

Interrupt: If enabled, this will prefix the input with "{{user}} interrupted you mid-sentence". Which can lead to unfortunate consequences...
Stop Stimulation: If enabled, this will instantly kill the current stim while sending the input.
Add To Prompt: If enabled, this will add a section to the system prompt. Useful to trigger specific behaviors or responses from the AI when this button is pressed.
Lines: A list of input lines from which the user's input will be randomly selected when pressing the quick reply button.

Delay Module

This module allows the AI to "freeze" the app for a certain amount of time. A visual counter is displayed, but its duration may be hidden (the AI decides). This can be useful to create anticipation or real timed tasks. For example: "Hold for 30 seconds", "I'm leaving, wait for me...", etc.
This kind of gives the AI a notion of time, and the ability to interact with it.

You can set the minimum duration for the delay.
You can set the maximum duration for the delay.
Usually the AI uses it once per response.

Special Effect Module

It took me a while to build, but I find that it adds a lot of depth and immersion to scenarios. I'll probably continue to tweak it and add more effects over time.
This is an agent that you can enable, which will analyze the AI's response and add special effects to the relevant segments.
It works very well with fantasy or sci-fi scenarios that usually involve a lot of visual effects.
I won't spoil the list of available effects, but I hope you'll like the added visual value!

This feature is only available to subscribers. As with the "Intelligent Image" and "Intelligent Stim" features, it requires a big, intelligent, and fast model.

Generation Enhancer

I wanted to add more spice to the AI's responses, make them go off-rail, be more proactive, take unexpected turns, emphasize certain aspects, etc. So I created this feature, to push their limits.
It's difficult to balance so it may feel a bit shy, or compeletely derail the plot into another direction sometimes. But overall I like the added unpredictability!
I recommend using a maximum of two traits at the same time. More will probably end up with conflicting instructions and a confused AI.

EstimAI Scenario & Editor improvements

I've added multiple new features to help you create even better experiences.

Scenario Editor:

Variables: You can now use the "Open value" type, which will replace the variable with a user-prompted value.
The State tracker has been improved: properties are now typed, and you can display a gauge using the 'range' type.
Scenario preset settings: you can recommend specific settings for your scenario to be played with. Then you can just click the "Apply recommended settings" in the Play page.
This can be useful if your scenario requires specific configurations (such as: MCQ mode, or Text Mode, etc.).

Trigger / Event:

New Trigger: Check State Data, since properties are now typed, you can easily check the state of specific properties.
New Event: Change the character image generation settings.
New Event: Change quick replies displayed.
New Event: Replace text in scenario (Regex compatible). This is extremely useful to create phases, without copying the complete scenario everytime.
New Event: Change EstimAI Settings: During a scenario you can trigger changes to the AI's settings. This has been useful for me for a scenario that requires setup at the start (done in Text Mode)then it switches to Cinematic Mode when the session really starts.

EstimAI Improvements:

Few important changes to EstimAI:

Intelligent Stimulation: The feature has been improved a lot, and should work a lot better now. I'm still gathering feedback!
Intelligent Image generation: Even more improvements, the best state it has ever been!
New Template: EudaimoniaV2. Written specifically with Eudaimonia's cinematic/visual novel aspect in mind.
Timestamps can now be enabled (giving AI the ability to track time more accurately)

Miscellaneous

Big improvements to the "Saved Sessions" page, with new features (duplicate, add message into history, etc.)
Creating / editing stimulations for Intiface is now possible: here.
A "Stop" button has been added to cancel the pending request in EstimAI.
New in-game settings.
Local TTS should now work properly, with support for the narrator's voice.
You can now upload images to illustrate your scenarios.
And a lot of other improvements and bug fixes!

What's next

Well, lots of things!
First, the Stereo E-stim rework, which is completely ready. But as I'm still unable to stim, I just can't test it properly. I'm very sorry for the wait.

Modules:
I'm very excited about this modular approach, and I've been thinking of multiple new modules to add. Here are a few that I'm starting to develop (not sure if they will see the light of day):
- Giving the AI the ability to use tools directly: dice, cards, sensors, QTEs, or anything that can be programmed client-side!
- Giving the AI the ability to "see you" and check what you're doing, (without sending real images of you in embarrassing situations) in real-time. I'm staying vague for now, as I'm still working on the implementation.
- Giving the AI biometric data about you, to monitor your physiological responses. This could be really useful to track arousal, edges, orgasms, etc. in real-time. (This one is more dystopian sci-fi level for now.)

New Game Mode:
I don't want to say too much about it, but I'd like to create a real simulation of a relationship (over a long period of time) with a Domme played by the AI.
Dedicated interface, easier setup, using the best of EstimAI and Eudaimonia's normal mode. Development will probably still take some time, but I hope this will be done by the end of the year.

I hope you'll like it, have fun!