Egomorph

EgoMorph is a browser-based emotional AI assistant designed to recognize feelings, adapt its behavior, and learn from user feedback.

The app runs as a Progressive Web App and can be installed on smartphones or desktops like a native app. Depending on the device, EgoMorph can run in different modes, from a lightweight Lite Mode to an API Mode with external AI integration.

EgoMorph can detect emotions such as joy, anger, sadness, fear, trust, and surprise. This allows the app to respond more personally and adjust its behavior based on the user’s mood. It also supports voice input, voice output, long-term memory, a thinking mode, and a feedback system that can improve emotion recognition over time.

A major focus of EgoMorph is privacy. In Lite, Standard, and Full Mode, all data stays locally in the browser. No server is required. In API Mode, users can connect EgoMorph to OpenAI-compatible endpoints such as Ollama, LM Studio, llama.cpp, OpenAI, OpenRouter, or other compatible backends.

EgoMorph offers four operating modes:

Lite Mode

Designed for weak devices and simple hosting. Uses keyword-based emotion recognition and requires almost no extra resources.

Standard Mode

Uses a local emotion model in the browser for better emotion recognition.

Full Mode

Combines a local emotion model with a local language model for more advanced replies.

API Mode

Connects EgoMorph to an external or local AI backend for stronger responses and flexible model selection.

Key features include emotion recognition, adaptive behavior, voice input and output, long-term memory, feedback-based learning, thinking mode, offline support after the first load, and flexible AI integration.

EgoMorph is made for users who want a personal, customizable, and experimental AI experience. It is open source, locally usable, and designed to be adapted by developers.