Hermes Agent Configuration Guide

A self-improving AI agent by Nous Research, supporting terminal interaction, multi-platform messaging (Telegram/Discord/Slack, etc.), skill learning, and memory systems.

Features

Feature	Description
Multi-model support	200+ models via OpenRouter, Anthropic, OpenAI, Gemini, etc.
Custom endpoints	Any OpenAI-compatible API
Terminal tools	Full terminal operations, file I/O, code execution
Messaging platforms	Telegram, Discord, Slack, WhatsApp, Signal
Skills system	Auto-creates and improves skills from experience
Memory system	Persistent cross-session memory and user profiling

Installation

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

After installation, reload your shell:

source ~/.bashrc  # or source ~/.zshrc

⚠️ Note

• Supported on Linux, macOS, WSL2, and Android (Termux) only
• Windows users must install WSL2 first

Configuring TokenSwitch

Hermes Agent supports custom OpenAI-compatible endpoints. You can connect to TokenSwitch via interactive setup or manual configuration.

Option 1: Interactive Setup (Recommended)

hermes model
# Select "Custom endpoint (self-hosted / VLLM / etc.)"
# API base URL: https://tokenswitch.cc/v1
# API key: your TokenSwitch token
# Model name: choose based on your needs

Option 2: Manual Configuration

Edit ~/.hermes/config.yaml:

model:
  default: your-model-name
  provider: custom
  base_url: https://tokenswitch.cc/v1
  api_key: your-tokenswitch-token

You can also store the API Key in ~/.hermes/.env:

OPENAI_API_KEY=your-tokenswitch-token

Configuration Reference

Setting	Value	Description
base_url	https://tokenswitch.cc/v1	TokenSwitch API endpoint
api_key	Your token	Obtained from the TokenSwitch platform
provider	custom	Use custom endpoint
default	Model name	Choose based on your token group's supported models

Verify Configuration

hermes

If the interactive interface launches and responds normally, the setup is complete.

FAQ

Model not responding

• Confirm base_url is https://tokenswitch.cc/v1
• Confirm the API Key is correct
• Confirm the token group supports the selected model

How to switch models

Use the /model command within a session to switch between configured models, or exit and run hermes model to add a new provider.

More Information

• GitHub Repository: NousResearch/hermes-agent
• Official Docs: hermes-agent.nousresearch.com/docs
• AI Providers: Providers Documentation