Agent OS Q&A: Setup, Loops + New AI Models
A Q&A session covering the Agent OS—a modular AI automation system where users can plug in various AI models and tools. The speaker demonstrates setup processes, discusses token optimization techniques, explains looping features in Claude Code, and addresses community questions about integrating multiple AI agents without overwhelming complexity.
Summary
The session begins with an overview of the Agent OS, a customizable system that allows users to integrate multiple AI agents and automate workflows within a single interface. The speaker highlights recent additions including a video agent built on Minimax H3, integration of Deep Seek V4 Flash (released 24 hours prior), and new Google Search Console integration for SEO content creation. Setup time is discussed as taking 20-30 minutes for experienced users and up to an hour for beginners, with API configuration adding 5-10 minutes.
The transcript covers several tools and services: Whisper Flow for voice-to-AI interaction with new note-taking features, Dina AI for 30-second video generation with multilingual support, and Buzz for agent workflow orchestration. A key concern raised is token usage and context bloat when chaining multiple agents; the speaker addresses this by recommending token minimization techniques like RTK Caveman and headroom ponytail, available through an open-source playbook in the AI Profit Boardroom.
For users struggling with tool complexity, the speaker recommends simplification—reducing to Claude Code alone rather than juggling Hermes, Buzz, and other tools simultaneously. Three looping approaches are explained: manual instruction ('keep running until finished'), the /goal command (loops until a judge model confirms goal completion), and the /loop command (time-interval based looping). The speaker notes spending 3-4 hours daily building and updating the system.
VPS deployment is addressed through case studies: Cloudflare tunnels for mobile/desktop access, manual agent-assisted setup on multiple devices, and Tailscale for cross-device access. The speaker uses Claude CLI locally rather than the API for development, leveraging existing context to quickly add features like new model integrations.
Key Insights
- The speaker claims that modular AI systems eliminate the feeling of being overwhelmed by rapid model releases because new models can be plugged in immediately without system overhaul, making the system better with each update rather than causing user anxiety.
- Setup time varies significantly by user experience level—20-30 minutes for experienced users, up to an hour for beginners, with most setup time spent on API configuration rather than the base system installation.
- Token bloat from multi-agent chaining is manageable through open-source token minimization techniques (RTK Caveman, headroom ponytail), and the speaker reports not approaching token limits despite running three separate workflows in one day.
- The speaker recommends simplifying to a single tool (Claude Code) with built-in looping features (/goal and /loop commands) rather than orchestrating multiple specialized tools, particularly for users finding the ecosystem confusing.
- The speaker maintains Agent OS locally for security reasons rather than on VPS, but notes community members have successfully deployed it via Cloudflare tunnels, Tailscale, and manual agent-assisted multi-device setup.
Topics
Transcript
[0:00] So today we're going to be answering the latest questions on the agent OS, which is a powerful system where you can plug all your agents in, build custom automations, have everything inside one tab, plug it into your memory, and basically build and automate anything that you want. To give you some examples over the last 24 hours, we have this video agent that is super powerful for, for example, adding avatar videos with uh fully edited video plugged in. As you can see, this was done literally with one single prompt, and it's on the new Miniax H3, which is pretty amazing in [0:30] itself. We also plugged in Deep Sea Coder. Deep Seek just dropped yesterday.…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Julian Goldie SEO
Google AI Studio Just Got A Huge Security Upgrade
Google is reportedly developing a new "Security Check" mode for AI Studio that would apply secure coding practices to AI-generated applications. The feature aims to automate security verification during app development, addressing common vulnerabilities like exposed keys, weak authentication, and data access issues before deployment.
New Gemini Skills Update Is WILD!
Google has launched Gemini Skills, a new feature allowing users to save custom instructions once and reuse them repeatedly by typing a slash command, replacing the discontinued Gems feature. Skills can include reference files, be combined together, and are built on an open file format (skill.md) that allows portability across different AI tools.
NEW Codex CLI Update Is ABSURD!
OpenAI's new Codex CLI update transforms terminal-based programming assistance by enabling multiple AI agents to work simultaneously on different tasks, visible in a full-screen control center interface. The update includes features like Agent View, work trees, voice control, and parallel task execution, fundamentally changing how AI assistance operates from sequential Q&A to autonomous team-based work management.
NEW Hermes Cloud + Agent OS is Crazy! 🤯
Hermes Cloud combined with Agent OS enables users to deploy AI agents in the cloud 24/7 using free models without API keys or local setup complexity. The system provides persistent memory, sandbox isolation, and integration with custom workflows like content research, competitor analysis, and news monitoring.
Manus 2.0 Just Changed AI Agents Forever
Manas 2.0 represents a fundamental architectural overhaul that enables users to create games, websites, videos, and automations through simple text requests, with the key innovation being editable components rather than fixed outputs. The platform uses a new Cascade engine that reduces token usage by 23% and improves task speed by 28%, while maintaining all project elements in one connected workspace.