Please note folks: version 0.7.6 is currently being finalized.

Vaudryllis is a local EVA — an evolving virtual entity / assistant designed around continuity, structured memory, voice/avatar presence, internal state, and long-term interaction with the user.

Vaudryllis is not a generic chatbot. It is designed as a local autonomous conversational system that can maintain continuity over time, learn progressively with its user, and evolve through interaction.

The current v0.7.5 build already includes the main Vaudryllis local experience: conversational interaction, persistent memory layers, avatar and voice-based presence, configurable runtime options, and a structured internal software brain inspired by biological organization.

Vaudryllis does not start as an all-knowing assistant with every topic or dictionary already built in. It is designed to learn progressively: by talking with the user, by listening to the desktop environment when configured, and by searching online information when needed. Over time, it becomes more adapted to its user, its context, and its own accumulated experience.

The next version, v0.7.6, is an important step in the project. It is planned to introduce a more complete EVA Control Center, giving users a clearer view of Vaudryllis’ internal state, modules, health, voice system, memory activity, runtime status, and overall progression.

V0.7.6 will also improve the linguistic base, response speed, voice interaction, and the general foundation required for Vaudryllis to move beyond a classic written interface.

This is the first step toward reducing the need for a traditional chatbox. Vaudryllis is being developed so that users can interact with her more naturally through voice and presence, while she becomes increasingly capable of speaking, writing, maintaining state, and taking small contextual initiatives.

The Founder/Beta Edition (Local) is intended for early adopters who want access to Vaudryllis in its founder-stage form while supporting its continued evolution toward a larger long-term vision.

Current public build: v0.7.5 Next planned update: v0.7.6

Key points:
• Local EVA / autonomous conversational system
• Persistent multi-layer memory
• Continuous internal state
• Voice and avatar-based presence
• Customizable avatar and voice
• Learns progressively with the user
• Can use conversation, desktop listening, and online research depending on configuration
• Structured software brain inspired by biological organization
• Designed for long continuity beyond short disposable conversations
• V0.7.6 will introduce a more complete EVA Control Center
• V0.7.6 will improve voice interaction, linguistic base, speed, and internal visibility
• First step toward interaction beyond a traditional chatbox

License policy:
• One-time purchase
• No heavy activation required
• Lifetime validity for the Vaudryllis Local line up to v0.9
• Future Founder/Beta Local builds are included up to v0.9
• Customers will receive update information by email when a new version is available
• Version 1.0+ is not included and will be treated as a separate future branch

Please note: Founder/Beta Edition includes updates for the Windows customer branch up to v0.9. V0.9 is planned as the complete Windows customer version for regular PC users. V1.0 is a separate future branch intended for a dedicated OS / robotic embodiment and may require specific hardware. Founder/Beta customers are not being asked to pay twice for the same Windows product.

Known issue (v0.7.5, non-French Windows systems) — 5-minute fix

- If Vaudryllis shows "translation model missing French to English" and doesn't answer typed messages:
1. Go to the official Argos model index: https://www.argosopentech.com/argospm/index/
2. Download Translate English ? French (translate-en_fr-….argosmodel) and Translate French ? English (translate-fr_en-….argosmodel).
3. Rename each file's extension from .argosmodel to .zip and extract them (one folder each).
4. Move both folders into: C:\Users\\.local\share\argos-translate\packages\ (paste the path into the File Explorer address bar; create the folders if they don't exist).
5. Restart Vaudryllis — typed conversation now works.

Note: voice input is a separate component that arrives in an upcoming update. The next version installs all language models automatically — this manual step will no longer be needed.