Skip to main content

Quick Overview

In this video, creator Alex Finn provides an exclusive walkthrough of the Hermes mobile application. He presents the tool's interface and demonstrates how it integrates with local desktop hardware and various artificial intelligence models. The video examines where Hermes fits in comparison to competing platforms like Grokbot and Claude Duet.

Key Points

  • 1.The Hermes mobile app functions as a remote interface connecting directly to a user's desktop Hermes instance rather than running agents in isolation.
  • 2.Unlike competing platforms that restrict each agent to a single ongoing conversation, Hermes supports multiple independent sessions under a single agent profile.
  • 3.Users can switch underlying models for any agent at will, selecting cloud providers or routing queries to local hardware setups such as a DGX Spark.
  • 4.Hermes automates cross-device machine management and file operations, requiring fewer permission prompts to execute commands and install software.
  • 5.Grokbot remains the preferred tool for cloud-based development and multi-bot delegation, while Claude Duet retains an edge in dedicated coding workflows.
  • 6.The Hermes mobile app is completely free and is scheduled for a broader public release.

Summary

Alex Finn introduces his early access walkthrough of the Hermes mobile application, highlighting how it integrates into an everyday AI workflow. Rather than operating as an isolated mobile assistant, the Hermes app connects across a network to the user's desktop setup, pulling in all desktop agents, sessions, and environment settings. The home screen displays running agents such as Hermes powered by GPT-6.1 Sol 900K and a local DeepSeek model running on a dual DGX Spark cluster. Finn shows how users can inspect live execution states, navigate active background tasks, and observe ongoing work across different machines.

Within the application interface, each bot offers comprehensive configuration options. Users can switch underlying foundational models on the fly, selecting options across providers such as OpenAI, Anthropic, and xAI, or selecting local instances. The settings allow fine-grained customization of individual skills, Model Context Protocol configurations, external plugins, automated routines, and active projects. Finn demonstrates auxiliary features including a mobile-friendly Kanban board and an Obsidian-style constellation map depicting the agent's interconnected memory nodes. In active chat threads, the app renders full technical logs of every tool call, terminal command, and web search executed on the host computer.

Finn outlines three main reasons why the Hermes mobile app stands out from alternatives like Grokbot, ChatGPT, and Claude. First, Hermes provides robust session management within each agent profile. While other tools limit an agent to a single ongoing conversation thread, Hermes allows dozens of distinct, concurrent sessions beneath a single persona. This allows a user on the go to log separate ideas, such as newsletter outlines or gaming configuration steps, without muddying a single conversation context. Second, Hermes is currently the only platform that cleanly bridges mobile access with local model execution, enabling unlimited local inference through personal hardware.

Third, Hermes excels at remote administrative control over physical computers. Finn recounts sending a command from his phone to install multiple AI applications and tools on a newly acquired MacBook Pro, as well as installing World of Warcraft add-ons on an RTX 5090 system. Hermes completed the tasks autonomously, running commands and verifying installations with minimal permission hurdles. Despite these strengths, Finn notes that Grokbot remains his preferred choice for complex knowledge work and multi-agent coordination due to its native cloud machine integration, while Claude Duet remains his primary tool for dedicated coding tasks with Claude Code.

App Interface and Configuration Architecture

The Hermes mobile application acts as a direct bridge to a user's desktop installation, displaying all configured agents and active tasks. Each agent contains deep configuration menus covering model selection, skills, plugins, Model Context Protocol connections, routines, and projects. The interface also supports interactive Kanban boards and memory constellation graphs showing linked agent memory. Within conversation threads, the app exposes transparent logs detailing command executions, tool calls, web queries, and elapsed execution times.

Parallel Sessions for Mobile Workflow

A primary advantage of the Hermes mobile app over platforms like Grokbot is its multi-session architecture per agent profile. Competing tools typically force all interactions with a specific agent into a single thread, leading to confused context when multiple distinct thoughts arise. Hermes permits users on the move to spawn separate task sessions under one agent, allowing simultaneous project tracking, content planning, and system administration without context pollution.

Local Model Support and Hardware Management

Hermes provides smooth access to local hardware setups, allowing users to run local models such as DeepSeek 4.1 Flash on dedicated hardware like a DGX Spark for unlimited usage without API restrictions. Furthermore, Hermes excels at managing local hardware remotely. Users can instruct the agent from a phone to set up software, configure environments, install add-ons, or run diagnostics across different desktop computers, executing tasks with minimal friction or repetitive permission prompts.

Comparative Roles Across the AI Agent Stack

While Hermes delivers strong utility for quick mobile administration, multi-session organization, and local execution, it does not entirely replace specialized tools. Grokbot remains superior for intensive knowledge work, cross-bot collaboration, and testing within dedicated cloud environments. Similarly, Claude Duet offers stronger capabilities for heavy programming due to its native integration with Claude Code.

The Bottom Line

The video demonstrates how the Hermes mobile app serves as an effective mobile control center for desktop agents and local machine automation. It establishes that Hermes solves major mobile limitations found in rival apps through multi-session tracking per agent, straightforward local model connectivity, and autonomous machine management. It concludes that while Hermes is an ideal daily driver for remote hardware tasks and rapid brainstorming, specialized platforms like Grokbot and Claude Duet remain necessary for deep cloud research and dedicated programming workflows.

FAQ

What is the Hermes mobile app and what are its primary features?

The Hermes mobile app is a mobile client that connects remotely to a user's desktop Hermes installation. It allows users to control AI agents, monitor commands and tool calls, run multiple conversation sessions per agent, switch models dynamically, and manage local computers from a smartphone.

How does session management in the Hermes mobile app differ from Grokbot?

Grokbot restricts each bot profile to a single ongoing conversation thread, meaning multiple distinct topics must share the same chat context. Hermes allows users to create and manage dozens of separate sessions under a single agent profile, preventing context confusion.

Can the Hermes mobile app connect to and run local artificial intelligence models?

Yes. The Hermes mobile app can connect to local hardware running private models, such as DeepSeek 4.1 Flash hosted locally on a DGX Spark, providing unlimited local usage without third-party API restrictions.

What administrative tasks can the Hermes mobile app perform on connected desktop hardware?

Hermes can execute terminal commands, manage system configurations, install desktop and mobile applications, search for and install game add-ons, and move files across connected computers with minimal permission prompts.

In what workflow scenarios does Alex Finn prefer using Grokbot over Hermes?

Finn prefers Grokbot for deep knowledge work and complex multi-agent delegation, as Grokbot provides isolated cloud computing environments to build and test projects before deployment.

Why does Alex Finn continue to use Claude Duet instead of Hermes for programming?

Finn uses Claude Duet for dedicated coding tasks because of its direct integration with Claude Code and the coding capabilities of Anthropic's Claude models.

Worth watching for

Developers, AI enthusiasts, and power users who run local models or multi-agent environments and want to manage workflows remotely from their mobile devices.

  • hermes
  • ai-agents
  • mobile-app
  • local-llm
  • grokbot
  • claude-duet