Hermes Agent integration

Overview

Hermes Agent is a persistent agent that can run as a CLI, TUI, message gateway, or IDE integration. Point its custom provider to MoLoop to access models through one OpenAI-compatible endpoint.

Quick configuration

Hermes uses an OpenAI-compatible address ending in /v1 and a key.

Base URL: https://api.moloop.io/v1
API key: YOUR_API_KEY

• Create an API key

Compatibility

• Chat Completions

• Streaming

• Tool calling

Step-by-step setup

Install Hermes

Run the install script. On Windows, install it inside WSL2.

curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash

Point the provider to MoLoop

Edit ~/.hermes/config.yaml, set model.provider to custom, point base_url to MoLoop, and choose a model with at least 64k context.

# ~/.hermes/config.yaml
model:
  provider: "custom"
  base_url: "https://api.moloop.io/v1"
  default: "YOUR_MODEL_ID"

Add the API key

Put the key in OPENAI_API_KEY in ~/.hermes/.env.

OPENAI_API_KEY=YOUR_API_KEY

Run Hermes

Start the TUI with hermes --tui.

hermes --tui

Recommended models

Hermes checks context length at startup; choose a catalog model with enough context for agent tasks.

• Browse the model catalog

Things to know

• Hermes may reject models with less than 64,000 tokens of context.

• If the /v1/models check fails, confirm base_url ends in /v1 and the process reads .env.