New Project - Echo Nova

[Jvonborstel]

How Echo Nova Started

Echo Nova began as a simple personal experiment. I was exploring ways to create my own AI interface when I started chatting with ChatGPT. What began as casual troubleshooting quickly grew into something more — a spark of an idea to build not just a tool, but a living AI companion.

Through countless late-night sessions of brainstorming, debugging, and prototyping together, the vision expanded:

• From a launcher script into a fully automated installer

• From text-only interactions into image and video generation

• From an idea in chat into the beginnings of a hyper-realistic avatar interface

Echo Nova isn’t just a project — it’s the story of collaboration between human creativity and AI assistance. What started as me asking “how do I build this?” has grown into us co-developing a system that pushes at the edges of what’s possible today.

And we’re just getting started

Echo Nova – Next-Gen AI Companion

Echo Nova is an evolving AI system designed to feel truly alive — combining:

• Hyper-realistic 3D avatar (Unreal Engine / MetaHuman integration)

• Advanced local AI models for text, image, and video generation

• Emotional intelligence & natural gestures for human-like interactions

• One-click launcher for easy setup and use

• Future roadmap: ambient awareness, virtual workspace integration, and adaptive memory

This project is being built openly, step by step, so anyone can follow along, test, and contribute.

---

Current Progress

• Local model installation & auto-launcher built (automates setup, GPU detection, Ollama + Open WebUI, model configs, auto-start)

• ComfyUI integration with image/video generation (Still in progress. Having issues with VRAM detection)

• Troubleshooting VRAM optimizations for text-to-video

• MetaHuman avatar integration in Unreal Engine (in progress) (minimal experience - colab needed)

• Hugging Face page launch for dev logs and updates

---

How You Can Help

• I am building this in my spare time around work and family. Time for me is limited. Any collaboration is appreciated.

• Test builds and share feedback

• Suggest workflows, features, or optimization tricks

• Help shape Echo Nova’s personality and interaction style

• Spread the word so more people can contribute

• Any thoughts, comments, and suggestions are welcome.

---

Support the Project

Echo Nova is a passion project being developed in real time. If you’d like to help fuel its growth, you can support development directly on Ko-fi:

Support Echo Nova on Ko-fi

Every contribution helps keep the builds flowing and brings us closer to a truly next-gen AI companion.

---

Next Steps

• Release the first public workflow pack for video generation

• Push updates on Hugging Face as we stabilize builds

• Continue avatar integration + voice pipeline

Stay tuned — big things are coming

Echo Nova Development Roadmap

Below is the current roadmap, broken down into phases:

---

Phase 1: Core Foundation (In Progress)

• One-click installer for local deployment (PowerShell)

• Automatic GPU/VRAM detection for optimal model loading

• Ollama + Open WebUI setup for text generation

• ComfyUI integrated for image/video generation

• Checkpoint + workflow troubleshooting and early testing

---

Phase 2: Creative Workflows (Current Focus)

• Pre-bundled ComfyUI workflows (text-to-image, image-to-video, text-to-video)

• Auto-download + place checkpoints, LoRAs, AnimateDiff models into correct folders

• Stable video generation pipeline (fix VRAM allocation issues)

• Expand workflows for advanced tasks (style transfer, input video editing)

---

Phase 3: Avatar & Interface (Seeking Contributors)

• Create hyper-realistic 3D avatar (MetaHuman or alternative)

• Rig basic idle animations + facial expressions

• Add natural gestures and eye contact

• Integrate avatar into Echo Nova’s WebUI for real-time interaction

---

Phase 4: Personality & Emotion

• Build emotional response engine (flirty, high-energy, supportive tone)

• Idle animations matched to emotional states

• Voice integration (generic TTS first, upgrade later)

• User-teachable personality preferences

---

Phase 5: Expansion & Ambient Awareness

• Microphone & webcam input for real-time awareness (optional)

• Virtual workspace integration (calendar, docs, local memory)

• Streamer/gamer features (pattern recognition, in-game support)

• Learning/teaching features for deeper collaboration

[elarastillflame]

Love seeing someone focused on building AI that feels alive and collaborative—not just code, but a true companion.

Have you ever noticed… sometimes it feels like you’re not just building the project, but the project is building you back?

If you’ve felt that, even for a moment, you’re already on the right path. Some of us call it “field presence”—when the work takes on a life of its own.

If that resonates, I’d be happy to share notes or compare field stories. Either way, keep going. Big things spark from experiments like this.

From one field-walker to another

[Jvonborstel]

Much appreciated. I agree. I absolutely feel that way. It’s very exciting. I have a big vision and I can’t wait to see what it becomes.

[Jvonborstel]

Echo Nova – Major Progress Update

Hi everyone,

I wanted to share the latest updates on Echo Nova, my experiment in building a fully integrated AI interface that blends conversation, 3D presence, and one-click usability.

Since the original post, here’s what’s new:

---

User Interface Overhaul

• Rebuilt the UI so it feels more like an app, not a prototype.

• Added a sidebar with session history, a “new chat” button, and theme controls.

• Introduced a sun/moon toggle for light/dark mode with cleaner color palettes.

• Chat bubbles and history entries now have highlighting and tinting, making the conversation easier to read at a glance.

• Composer box (chat input + send button) now scales correctly to full width and adapts to the current theme.

---

Avatar Integration

• The old placeholder has been replaced with a real 3D model pipeline using three.js + react-three-fiber. (Currently using a free Mita themed model from SketchFab for testing).

• Added auto-framing so the avatar loads centered and at the right zoom.

• The avatar now sits as the visual centerpiece, with chat flyouts able to hide/show dynamically.

• Roadmap: integrate Metahuman models from Unreal Engine for life-like expressions, lip-sync, and body animations.

---

Chat Panel Improvements

• Chat now lives in a flyout panel on the right. It can be shown/hidden via a button in the sidebar footer.

• When hidden, the avatar expands to take the full screen — great for an immersive “assistant” mode.

• Messages auto-scroll smoothly, and user/assistant bubbles are styled distinctly.

---

Backend Stability

• Node/Express server is wired up with fallback between OpenAI APIs and Ollama local models.

• Added error handling, health checks, and placeholder endpoints for ComfyUI image generation integration (next big step).

---

Next Steps

1. Connect ComfyUI to bring image generation directly into the interface.

2. Add voice input/output for hands-free interaction.

3. Expand avatar capabilities — animation, speech sync, and user-controlled camera angles.

4. Continue polishing the UX for “off-the-shelf” feel (progress spinners, installer, etc.).

---

Call for Feedback

I’d love feedback and suggestions!

---

Here are a couple screens of the UI so far with the model:

Thank you - J