🚀 Agentic AI with Hermes Agent starts Fri, Sep 25 — spots are limitedRegister Nowor Email jayaram.linux@gmail.com
SaturdAI.
← Agentic AI with Hermes Agent

Day 2 · Mon, Sep 28, 2026

Hermes Fundamentals

Two panels: a Day 1 recap checklist on the left, and today's goal — get Hermes talking to a real local LLM end to end — on the right

Day 1 was a talk: what an LLM is, why Hermes Agent is provider-agnostic, a first look at the DGX Spark and Mac Mini, and the shape of the next six sessions. Day 2 is where the hands-on labs actually start. Today's goal is narrow and concrete: point Hermes at a local model, send it a real prompt, and confirm the whole chain works with a health check — nothing more advanced than that yet.

1 / 20

🎥 Watch the recording: https://www.youtube.com/live/jyevJPcrjMM

Hermes Agentis our vehicle for this course because it doesn't need a hosted API key — it talks to any OpenAI-compatible endpoint, which means it works against a local llama-server instance on our DGX Spark or Sanjay's Mac Mini exactly the same way it would against a cloud model. Today is about getting comfortable with what Hermes is and getting it talking to a local LLM end to end.

Objectives

  • Understand what Hermes Agent is and why it's provider-agnostic
  • Point Hermes at a local, OpenAI-compatible endpoint instead of a cloud provider
  • Run a first one-shot prompt and confirm the round trip actually works
  • Run hermes doctor to sanity-check the setup before building on it

Lab: point Hermes at a local model

Configure Hermes with a custom provider pointing at the local endpoint we'll be running on the DGX Spark:

$ hermes config

Then run a single prompt to completion with --clito confirm it's actually talking to our local model and not a cloud fallback:

$ hermes -z "In one sentence, who are you and what model are you running on?" --cli

Finally, run the built-in health check — this is the fastest way to catch a broken endpoint, missing dependency, or misconfigured chat template before it wastes time later in the course:

$ hermes doctor

Session notes: to be filled in after the live session on Sep 28, 2026 — actual hardware used (DGX Spark vs. Mac Mini), model loaded, and real command output.