One Chat App for All My Agents: Open WebUI
May 18, 2026•Channel
AI Analysis
Data from YouTube Data API v3•Updated Just now
Video Overview
Video Details
Published1 week ago
Duration1:21:40
Video IDNHvcK-dnhAE
Languageen
CategoryPeople & Blogs
PrivacyPublic
Made for KidsNo
Video TypeRegular Video
Performance Metrics
Views1.7K
Likes52
Comments2
Engagement Rate3.23%
Likes per 100 views3.11
Comments per 1K views1.20
Video Tags
#vibe coding#vibe coding studio#youtube studio setup#sony fx30#fx30 setup#fx30 youtube#obs studio#obs setup#claude code#claude code tutorial#ai coding#ai studio#youtube studio tour#content creator setup#creator studio#youtube live#live stream setup#studio rebuild#home studio setup#youtube camera setup
Description
👑 Premium Community: https://mrc.fm/premium
I've ended up with AI agents all over the place. Hermes looks after my health data, my OpenClaw crew runs my socials, my calendar and a stack of automations, and I've got local models humming away on a Mac Studio. The problem? They all live in different apps, terminals and dashboards.
So in this stream I'm fixing that. I'm setting up Open WebUI as one clean chat app that ties the whole lot together. I link it straight to my Hermes agent and my OpenClaw agents, point it at my own local models (a 32B Qwen and GPT OSS 120B running on the Mac Studio, zero cloud), turn it into a proper app on my phone, and even give it a real terminal so a model can run code and build things for me.
By the end I've got one place to talk to every agent, fire off automations, check my calendar and poke at my home network. Built live, on my own kit, start to finish. Come hang out and build it with me.
00:00 Intro: Unifying scattered AI agents with Open WebUI
01:22 Testing local models (Qwen 3.6 & GPT-OSS)
04:35 Switching to Hermes agent & persistent chat threads
06:24 Running Open WebUI on iPhone (PWA & dictate mode)
13:31 Attempting web search & using Claude Code to set it up
21:52 Exploring calendars and scheduled automations
28:58 Success: Local model performing a live web search
31:31 Updating the AI agent's long-term memory with search results
34:52 Setting up daily web-scraping automations
36:29 Open Terminal: Running Claude Code inside the browser
38:11 Autonomous coding: Making a Python space shooter with Qwen
45:30 Installing DeepSeek V4 locally
49:36 DeepSeek voice dictation & live web search tests
56:04 Integrating ElevenLabs for Text-to-Speech (TTS)
1:00:15 Studio upgrade: How AI leveled up my YouTube setup (LinkedIn article)
1:08:14 Troubleshooting the ElevenLabs API integration
1:14:48 Admin panel: Analytics, chat evaluations & model split-testing
1:18:25 Success: ElevenLabs Text-to-Speech (TTS) works!
1:19:10 Live mobile voice chat demonstration
1:20:25 Outro & ASCII art goodbye