Building Agentic AI solutions in enterprise environments✦Training for IronMan 2027✦Building Aether Ops · AI-era governance for small, mid & large enterprise✦Ultra Marathon 2028 · because ERP wasn't hard enough✦Leading system integration post-$13B IPG acquisition at Omnicom✦Finance nerd by day. AI builder by night. Endurance athlete on weekends.✦8+ years · $17B+ in transactions · 150+ global business units✦Currently shipping: Aether Ops · Onyx · and things I can't talk about yet✦Stevens Institute · MS Business Intelligence & Data Analytics✦Ask my agent anything 👇✦Building Agentic AI solutions in enterprise environments✦Training for IronMan 2027✦Building Aether Ops · AI-era governance for small, mid & large enterprise✦Ultra Marathon 2028 · because ERP wasn't hard enough✦Leading system integration post-$13B IPG acquisition at Omnicom✦Finance nerd by day. AI builder by night. Endurance athlete on weekends.✦8+ years · $17B+ in transactions · 150+ global business units✦Currently shipping: Aether Ops · Onyx · and things I can't talk about yet✦Stevens Institute · MS Business Intelligence & Data Analytics✦Ask my agent anything 👇✦
HomeAboutNotesBuildsStack
Schedule↗

Govind Waghmare

Agentic AI for ERP. Cutting through red tape. Building in public from NYC.

govindwaghmare@icloud.comcal.com/govindw

Pages

HomeNotesBuildsAboutStackPartners

Social

XLinkedInGitHub
© 2026 · New York City·RSS·llms.txt

Hermes Agent: What It Is and How to Set It Up

Apr 15, 2026

Hermes is a family of fine-tuned models from NousResearch, built specifically for agentic and tool-use tasks. If you've been frustrated by base models that lose track of instructions mid-task, misfire on function calls, or refuse to follow complex multi-step workflows, Hermes is worth your time.

What makes Hermes different:

Base models like GPT-4o and Claude are trained for broad helpfulness. They're good at conversation. Hermes is fine-tuned on instruction-following data, tool use, and structured output, which means it stays on task longer, follows complex system prompts more faithfully, and handles function calling with less slop.

For production agentic workflows, the kind where your agent is making decisions across multiple steps with real consequences, this matters.

The models:

  • Hermes 3 (latest), Based on Llama 3.1/3.2. Best for complex instruction following and tool use.
  • Hermes 2 Pro, Strong structured output and JSON mode. Good for data extraction pipelines.
  • Both are open-source and available on HuggingFace.

Setup via Ollama (local):

# Install Ollama
curl -fsSL https://ollama.ai/install.sh | sh

# Pull Hermes 3
ollama pull nous-hermes3

# Run
ollama run nous-hermes3

Setup via OpenRouter (hosted):

# In your app, point to OpenRouter
curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -d '{
    "model": "nousresearch/hermes-3-llama-3.1-70b",
    "messages": [{"role": "user", "content": "Your prompt here"}]
  }'

Where I use it:

In multi-step ERP automation workflows where the agent needs to follow a structured procedure, pull data from one system, validate it, transform it, post to another, Hermes holds the instruction set better than base models. Less drift over long context windows.

It's not replacing Claude or GPT for general reasoning. It's a specialist. Use it where instruction adherence is the bottleneck.

The honest take: Most people reach for GPT-4o or Claude for everything. For agentic pipelines with well-defined procedures, Hermes is often cheaper, faster, and more reliable. Worth having in the toolkit.

Ask a follow-up about this note
$