Hermes Agent - Desktop

It is impossible to stay on top of all the new stuff. but this one is interesting as an interim solution - as I prefer Open WebUI, but I need to learn more how to get out of the sandbox. Hermes is not so picky about security ...Β 

Let's give it a try on my private laptop - using our M4 as the 🧠,

$ curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
$ How would you like to set up Hermes?
  ↑↓ navigate  ENTER/SPACE select  ESC cancel

   (●) Quick Setup (Nous Portal) β€” free OAuth login, no API keys, model + tools (recommended)
   (β—‹) Full setup β€” configure every provider, tool & option yourself (bring your own keys)
 β†’ (β—‹) Blank Slate β€” everything off except the bare minimum; opt in to each capability

$ Select provider:
  ↑↓ navigate  ENTER/SPACE select  ESC cancel

   (β—‹) Google AI Studio (Native Gemini API)
   (β—‹) DeepSeek (V3, R1, coder, direct API)
   (β—‹) Z.AI / GLM (Zhipu direct API)
   (β—‹) Kimi / Moonshot β–Έ (Coding Plan, Moonshot global & China endpoints)
   (β—‹) StepFun Step Plan (Agent / coding models via Step Plan API)
   (β—‹) MiniMax β–Έ (Global, OAuth Coding Plan & China endpoints)
   (β—‹) Ollama Cloud (Cloud-hosted open models, ollama.com)
   (β—‹) Arcee AI (Trinity models, direct API)
   (β—‹) GMI Cloud (Multi-model direct API)
   (β—‹) Kilo Code (Kilo Gateway API)
   (β—‹) OpenCode β–Έ (Zen pay-as-you-go or Go subscription)
   (β—‹) AWS Bedrock (Claude, Nova, Llama, DeepSeek; IAM or API key)
   (β—‹) Azure Foundry (OpenAI-style or Anthropic-style endpoint, your Azure AI deployment)
   (β—‹) Qwen OAuth (Reuses local Qwen CLI login)
   (β—‹) Alibaba Cloud Coding Plan (Dedicated coding tier)
 β†’ (β—‹) custom (direct API)
   (β—‹) Custom endpoint (enter URL manually)
   (β—‹) Configure auxiliary models...
   (β—‹) Leave unchanged
# ... some private stuff
Select terminal backend:
  ↑↓ navigate  ENTER/SPACE select  ESC cancel

   (β—‹) Local - run directly on this machine (default)
   (β—‹) Docker - isolated container with configurable resources
   (β—‹) Modal - serverless cloud sandbox
   (β—‹) SSH - run on a remote machine
   (β—‹) Daytona - persistent cloud development environment
   (β—‹) Singularity/Apptainer - HPC-friendly container
 β†’ (●) Keep current (local)

Your minimal agent is ready. What next?
  ↑↓ navigate  ENTER/SPACE select  ESC cancel

 β†’ (●) Start with everything disabled β€” finish now (most minimal)
   (β—‹) Walk through all configurations β€” opt in to tools, skills, plugins, MCP

$ source ~/.bashrc
$ volta run --node 22 hermes desktop
β†’ Installing desktop workspace dependencies...

Let's ask it the questions of all questions to see if it works ...

Yep - you're right πŸ‘

Links

Related post

Projects

The Meta-Agent

AI development is not only about maximizing generative output; it is also about maximizing reliability, controllability and costs. Frontier models have excellent reasoning abilities and a huge knowledge. They are brilliant minds which can handle most problems autonomously.

But there's a price to pay (literally in πŸ’²), as well as privacy risks and infrastructure dependency.

The S.H.O.D.A.N. project is for my personal education - but also to mitigate those risks without loosing quality. Mainly driven by my personal token consumption - which is way higher than I'm willing to pay for πŸ˜‰.

Comments