everyone is building jarvis & friday i'm building ultron

@luke.builds.ai

39.2K followers · usually 23.7K views a post

Views

347.0K

Likes

30.0K

Comments

768

Shares

1.8K

SoundTikTok sound · original sound - luke.builds.ai

Promoted product

Azaris AIhttps://lukebuildsai.com ↗

Why it worked

The video effectively uses a 'build in public' narrative, showcasing a complex AI project with a clear demonstration of its components and functionality. The comparison to popular AI assistants like Jarvis and Friday, coupled with the intriguing mention of 'Ultron', creates intrigue and positions the project as ambitious and cutting-edge.

Summary

The video demonstrates how to build an advanced AI agent, contrasting it with simpler AI assistants. It showcases the use of tools like Next.js, Claude Design, AirTouch, Hermes, and ElevenLabs to create a sophisticated AI system controlled by voice commands and hand gestures.

Structure

1Introduction to building advanced AI agents
2Demonstration of Next.js app setup
3Showcasing UI design with Claude Design
4Explaining finger tracking with AirTouch
5Detailing the use of Hermes for LLM integration
6Explaining voice control with Hermes Voice and ElevenLabs

Original caption

Breaking down how this setup actually works. There are a few different tools doing the work here, and some of them are pretty interesting. Original video by @sagar_builds #ai #aiagents #automation

More from @luke.builds.ai

Product placement

Jarvis & Friday (fictional AI assistants): appear in the video as examples of common AI builds. Removing them would not change what happens in the video. Ultron (fictional AI): appears in the video as an example of a more advanced AI build. Removing it would not change what happens in the video. Next.js (web framework): is used to start up an app. Removing it would change what happens in the video. Claude Design (UI/UX design tool): is used to build out a UI. Removing it would change what happens in the video. AirTouch (tool): is used for finger tracking, allowing cameras around a room or the webcam to track hands. Removing it would change what happens in the video. Hermes (AI agent framework): is used to connect an LLM to a computer and give it access to tools. Removing it would change what happens in the video. Python automation (framework): is assumed to be used to connect to phones and ask AI to turn them on/off. Removing it would change what happens in the video. Hermes Voice (AI voice): is used to make the AI agent sound a certain way. Removing it would change what happens in the video. ElevenLabs (AI voice platform): can be used to add a voice on top of Hermes Voice. Removing it would change what happens in the video. Azaris.ai (company): Creator is cofounder. Appears in creator bio. Removing it would not change what happens in the video. lukebuildsai.com (website): Creator's website, leads to a setup guide waitlist. Appears in creator bio. Removing it would not change what happens in the video.

Call to action

Comment "Jarvis" to get on how you can set up just like this for yourself.

On-screen text

everyone is building jarvis & friday i'm building ultron Claude Design I'm your three agentic harness like Hermes, you can connect to a computer and give it access to tools. I'm assuming he's using automation to connect to these phones to ask his AI to turn them on and off. For that voice, he's using Hermes Voice to be able Hermes agent. If you want it a specific way, you can always add an ElevenLabs voice on top of it a certain way, and then you can control your agentic harness using voice commands like you see here. Comment "Jarvis" to get on how you can set up just like this for yourself.

Transcript

So that command he just ran was to start up a Next.js app, which pulls up the screen that you see next. For that UI, he used Claude Design to build out that UI. For the finger tracking, he's using a tool called AirTouch. It allows you to put cameras around in your room or the web camera in front of him to move his hands around with that UI. You can unlock all three devices. Using an agentic harness like Hermes, you can connect an LLM to a computer and give it access to tools. I'm assuming he's using some Python automation to connect to these phones to ask his AI to turn them on and off. For that voice, he's using Hermes Voice to be able to be Hermes agent. If you want it a specific way, you can always add an ElevenLabs voice on top of it a certain way, and then you can control your agentic harness using voice commands like you see here. Comment "Jarvis" to get on how you can set up just like this for yourself.