Hermes Agent

Hermes Agent is the CLI from Nous Research, not the Hermes model family. Point a custom OpenAI-compatible endpoint at orientx and set api_mode to Chat Completions.


Base URL	`https://api.orientx.ai/v1`
Auth	`OPENAI_API_KEY` in `~/.hermes/.env`
Config	`~/.hermes/config.yaml`
Official docs	AI Providers

Install

curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash

On first launch choose Full setup (not Nous Portal) if you want to enter the endpoint there. Otherwise install, then run hermes model.

Point it at orientx

hermes config set OPENAI_API_KEY "$ORIENTX_API_KEY"
hermes config set OPENAI_BASE_URL https://api.orientx.ai/v1
hermes config set model <MODEL_ID>

Or write the files yourself. ~/.hermes/.env:

OPENAI_API_KEY=sk-...
OPENAI_BASE_URL=https://api.orientx.ai/v1

~/.hermes/config.yaml:

model:
  provider: custom
  default: <MODEL_ID>
  base_url: https://api.orientx.ai/v1
  api_mode: chat_completions

api_mode: chat_completions is required. Without it Hermes may wait for a Responses-shaped reply and never finish.

Run Hermes

hermes

Or hermes --tui. Switch models later with hermes model.

Hermes wants a large context window (64k+). If startup fails on context length, pick another id from GET /v1/models.

Install

Point it at orientx

Run Hermes

On this page