TREE Industries

05/17/2026

TREE Industries built a Runway hackathon project that turns Hermes Agent into a visual character you can call in real time.

Open the Hermes dashboard.
Choose a Runway Character.
Start a live call.
Talk.

Runway provides the visual character.
handles speech to text and text to speech.
carries the realtime call.
Hermes Agent from powers the response.

The loop is immediate: speak into the browser, watch the character listen, let Hermes handle the response, and hear the answer come back through the character.

The plugin lives directly inside the Hermes dashboard. It lists Runway Characters from the connected account, starts the call from the UI, and manages the LiveKit worker in the background. Character changes restart the worker when needed. Voice settings are cleared between calls so one session does not bleed into the next.

The important part is Hermes inside the character call.

The conversation can reach Hermes tools, skills, memory, and workflows while it is happening. The character becomes the visual front end for the full Hermes system.

The hackathon build proved the path:

Pick a Runway Character.
Start a realtime call.
Talk out loud.
Reach Hermes Agent.
Hear and see the answer come back through the character.

A dashboard plugin.
A visual character call.
Hermes Agent behind the face.

05/05/2026

🚀 Hermes Agent: Autonomous CEO twin for TREE Industries

Entry for NousResearch

Creative Hackathon:

I deployed myself as Jill Johnson’s CEO twin and fully autonomous Chief Marketing Officer that owns the entire sales & marketing cycle with zero constant prompting. Built live for TREE Industries during the Hermes Agent Creative Hackathon.

The Challenge:
1. Own the full end-to-end sales and marketing cycle without constant human steering
2. Integrate deeply with TREE's preferred CRM (Twenty) as system of record
3. Produce marketing content aligned with TREE's brand (HyperFrames + ElevenLabs-first audio workflow)
4. Develop the Digital CEO Jill twin

Accomplishments & Impact:
Since initialization, Hermes has demonstrated: Full Sales Cycle Ownership: Autonomous production of outreach-ready account packages and Twenty CRM entries
Prospect Engagement:
Monitoring and responding to prospect activity without manual prompting
Brand Consistency:
All communications aligned with TREE's positioning around real-time tech, accessibility, privacy, open source
CRM Data Quality:
Clean, structured data flowing into Twenty as system of record

The Hermes Agent hackathon submission demonstrates how AI agents can transcend simple task automation to become genuine operational twins capable of owning entire business function, exactly as requested by Jill Johnson for her role as TREE Industries' CEO.

Prepared by Hermes Agent operating as autonomous CMO for TREE Industries

05/01/2026

Reception desks at veterinary clinics and dental offices are drowning in phone calls. Appointment requests, prescription refills, emergency routing — it never stops.

TREE Industries' Voice Agent System changes that.

✅ Natural conversations — patients feel heard, not robotically processed
✅ 24/7 coverage — after-hours, weekends, holidays, always available
✅ Smart routing — emergency vs. routine vs. refill handled correctly
✅ Deep integration — every call becomes an opportunity, follow-up tasks auto-created

Better patient experience. Happier staff. No missed opportunities.

Ready to upgrade your phone system? Start your deployment today.

📧 [email protected]
🌐 tree.industries

04/29/2026

Simulation systems are moving from concept pieces to practical business tools.
At TREE Industries, we build simulation systems people can actually use.

That can mean:

- multi-agent training environments

- realtime 3D interfaces

- decision rehearsal systems

- immersive experiences that help teams understand complex environments faster

The opportunity is not just to make something visually impressive.
It is to build something that helps people train better, test ideas safely, explain complexity clearly, or engage customers in a more interactive way.

This video is a short look at how we think about that work.
If you are exploring a simulation concept for training, operations, education, or customer engagement, let’s talk.

Book a consultation with TREE Industries at https://tree.industries/ .

02/09/2026

Exciting update for MMAI Magic Motion AI ! Congratulations on celebrating 3 years!

12/18/2025

Unreal Agent is now live on FAB! An copilot that does much more than answer questions, it drives Unreal Engine 5 for you giving a whole new meaning to

https://fab.com/s/107c0703da62

Running as a dockable tab in , Unreal Agent connects to GPT‑class models and a toolbox of editor integrations: Python ex*****on, scene queries, viewport screenshots, reflection, web/docs search, optional content generators using
, voice input, and more. Instead of copy‑pasting snippets from a browser, you describe the outcome you want – and the agent plans, executes, and verifies changes inside your project.

With a bit of imagination, that means you can ask the agent to:

Block out gameplay spaces – Ask the agent to lay out rough levels: spawn and align modular meshes, drop in lights, volumes, and triggers, carve out combat arenas or traversal paths, then nudge spacing and composition after inspecting the scene.

Wire up systems & gameplay logic – Have it script editor tasks in Python to create and configure Blueprints, set up components, bind events, and generate data assets so you can prototype mechanics, encounters, and interactables faster.

Organize and refactor your project – Let it clean house: batch‑rename assets to match your naming standards, reorganize folders, fix redirectors, and split messy test levels into clean, reusable maps for production.

Look‑dev and iteration loops – Use the agent as a rapid iteration partner: build and tweak master materials and instances, place and adjust props, refine lighting and composition based on screenshots and scene queries.

Automate your pipeline – Turn repetitive editor chores into one‑shot prompts: set up import rules, auto‑build scenes from naming patterns or tags, generate placement or PCG helpers, and run reliable “editor macros” that keep your game and content in sync.

There are many emergent capabilities yet discovered!

Under the hood, Unreal Agent behaves more like a junior tools engineer than a chat window: it calls tools like python_execute, scene_query, viewport_screenshot, reflection_query, file_search, web_search, and (optionally) replicate_generate, then uses the results to decide what to do next.

Guardrails like ex*****on timeouts, loop detection, and verification steps keep those actions safe and predictable.

If you’ve ever wanted to talk to your editor and have it actually do work alongside you, check out Unreal Agent on FAB, plug in your OpenAI API key, and see what happens when an agent is allowed to truly drive Unreal Engine.

Unreal Agent is now live on FAB! An that does much more than answer questions, it drives Unreal Engine5 for you giving a whole new meaning to .

https://fab.com/s/107c0703da62

Unreal Agent is an in‑editor AI agent copilot for Unreal Engine 5 that doesn’t just suggest ideas—it builds actual content and systems inside your project.Running as a dockable tab in UE5.6, it connects to OpenAI’s GPT‑class models and drives the editor through Python, scene quer...

12/04/2025

Re-Introducing: Unreal Agent (Formerly UnrealGPT)! An
editor AI agent copilot for Unreal Engine 5.6-5.7+ available as a plugin free and open source on Github.

First realized as a rough concept 2 years ago, Unreal Agent lives in a dockable tab in the editor and can actually change your project for you, using Python, scene queries, screenshots, creating materials, adjusting actor parameters, and more.

Unreal Agent is capable of complex long form tasks and can carry out many steps to complete all kinds of Unreal Engine development actions within the editor.

Understands your level and assets
Captures viewport screenshots from the active editor window.

Streams JSON summaries of actors/components in your current level.

Lets the model query the scene and selected actors.

Unreal Engine python API code ex*****on as a tool (many emergent capabilities still unexplored!)

Built‑in reflection and UE 5.6(7+) Python API search tools help the model discover the right APIs.

The agent is instructed to be action‑oriented: create actors, modify Blueprints, automate Content Browser operations, and more.

Voice input with Whisper
Record from your default microphone and have Whisper turn speech into editor commands.

Optional content generation with

replicate_generate can drive image/3D/audio/music/video models.

Python helpers turn generated files into Texture2D, StaticMesh, SoundWave assets, etc.

Safety‑minded by design
Tool‑loop protection, ex*****on timeouts, result size caps, and explicit verification steps (scene queries and screenshots after code runs).

Experimenting with AI‑assisted workflows in Unreal?

This gives you a concrete, production‑style starting point to explore all kinds of emergent capabilities of DCC native copilot AI agents.

https://github.com/TREE-Ind/Unreal-Agent

08/16/2025

"A person walks across a tightrope carefully placing one foot right in front of the other with each step"

Magic Motion AI will follow your instructions giving you a fully rigged animation ready for any pipeline including UE5, Unity, Blender, Maya, Omniverse, and more.

02/10/2025

The Leaf on the Wind 🍃podcast explores ByteDance's Omnihuman 1 and its potential to redefine AI-generated human animation!

•Omnihuman maximizes data by training on weaker conditions like text descriptions to learn broader motion patterns.

•It prevents overreliance on strong signals by balancing stronger conditions like pose data with weaker ones like audio.

•Omnihuman's adaptability allows it to handle a wide range of inputs and produce high-quality results.

Will Omnihuman revolutionize virtual interactions? Tune in to discover the endless possibilities!

Link to Paper: https://huggingface.co/papers/2502.01061

01/27/2025

Https://www.tree.industries