Use Hermes Agent with RunBridge AI

Hermes Agent connects to RunBridge AI through a custom provider. This guide uses the OpenAI-compatible Chat Completions API.

Prerequisites

macOS, Linux, or WSL2, with the prerequisites in the official installation guide.
An active API key from the RunBridge AI console.
A text model available to your key that supports Chat Completions, streaming, and tool calling. Replace your-model-id throughout the example.

Install Hermes Agent

Follow the official installer instructions, then confirm the CLI is available:

hermes version

Configure the provider

Add your key to the user-level ~/.hermes/.env file. Edit an existing entry instead of adding duplicates, and keep this file outside your project repository:

RUNBRIDGE_API_KEY=your-runbridge-api-key

Merge the following into ~/.hermes/config.yaml, preserving any other providers and settings:

model:
  provider: custom:runbridge
  default: your-model-id

providers:
  runbridge:
    base_url: https://api.runbridge.ai/v1
    key_env: RUNBRIDGE_API_KEY
    transport: chat_completions
    default_model: your-model-id

runbridge is a name you assign to this custom provider. key_env tells Hermes which environment variable holds its key; transport selects Chat Completions. Keep /v1 in the base URL. Hermes adds /chat/completions and sends the key as a Bearer token.

Hermes can use separate auxiliary models for tasks such as vision and summarization. Review any existing auxiliary overrides if those tasks should also use RunBridge AI.

Verify the connection

Check the configuration, then start a new session so it reads the selected provider:

hermes config check
hermes chat -q "Reply with CONNECTED only."

The chat command makes a real API request and may incur usage. Confirm that it returns a response without authentication or endpoint errors. Then try a small coding task in a test project to check tool calling; a text response alone does not verify every agent capability.

Troubleshooting

Symptom	Check
Missing key or HTTP 401	Confirm `RUNBRIDGE_API_KEY` exists in `~/.hermes/.env` and `providers.runbridge.key_env` matches its name. Restart Hermes after editing it.
HTTP 404	Use `https://api.runbridge.ai/v1` and `transport: chat_completions`; do not append `/chat/completions` yourself.
Model not found	Use an exact model ID supported by your RunBridge AI key and this API format.
Existing chat uses the previous provider	Start a new session, or use Hermes’s `/model` command to select the configured provider.
Tool calls fail or request parameters are rejected	Check the model’s tool support and accepted reasoning settings. A model accepting ordinary chat is not necessarily suitable for an agent.

See the official provider configuration and model configuration references for additional options.

Get started

Coding agents

Agent platforms

Editor extensions

Use Hermes Agent with RunBridge AI

Prerequisites

Install Hermes Agent

Configure the provider

Verify the connection

Troubleshooting

​Prerequisites

​Install Hermes Agent

​Configure the provider

​Verify the connection

​Troubleshooting

Prerequisites

Install Hermes Agent

Configure the provider

Verify the connection

Troubleshooting