DESKTOP · EARLY SUPPORT

Use Hermes Agent with a local LLM on Mac

Try Hermes Agent with a local LLM on Mac through Bridge1. Early support covers a one-turn smoke test; configure the custom endpoint and key.

Updated 2026-10-29

Hermes Agent has early support: a one-turn smoke test only. Setup may change.

A featureless clay cyclist carries an orange satchel across a faceted stone bridge.

Before connecting Hermes Agent

For Hermes Agent, start with an Apple Silicon Mac: 24 GB minimum, 32 GB or more recommended. Check memory fit.
Bridge1 lists Ornith 1.5 35B-A3B (OptiQ 4-bit · default for coding) as its default for coding. Check the model's memory and context before choosing it for Hermes Agent.
Install Bridge1 through invite-only macOS early access, choose a model and open the Gateway panel for Hermes Agent.

Hermes Agent API and model names

Hermes Agent uses a custom OpenAI-format endpoint ending in /v1.

Set Default Model to route1 to follow your Bridge1 selection.

How Route1 chooses a model →

Hermes has no Bridge1 launcher; use the manual setup below.

Set up Hermes Agent

Use Gateway’s custom-endpoint values. This example shows Route1 and one onboarded model; your Gateway panel lists your own models.

Open Settings → Providers → Custom Endpoints → Add Endpoint.
Enter http://127.0.0.1:54321/v1 as the endpoint, your non-blank gateway key and route1 as Default Model.
Leave Discover models checked. Click Test, then Save.
Copy Gateway’s YAML into ~/.hermes/config.yaml under your endpoint: extra_headers → X-Bridge1-Client: hermes. Keep all models entries, supports_vision and context_length when supplied. Copy it again after onboarding another model.

providers:
  bridge1:
    extra_headers:
      X-Bridge1-Client: hermes
    models:
      "route1":
        supports_vision: true
        context_length: 262144
      "mtp-qwen3.6-35b-a3b-gguf":
        supports_vision: true
        context_length: 262144

First connection check

In Hermes Agent, type bridge1: help in a chat to see status, list models, or load, switch or stop a model. If a model is not ready, Bridge1 explains the next step in the chat. You can turn chat controls off in Settings.

Check that it works

Confirm Hermes uses the Bridge1 custom endpoint and chosen plain model id. Use Test, then Save; success here establishes the connection.
Ask an ordinary question in Hermes. A relevant reply completes the published one-turn smoke check.
A disposable-project task check is not part of Hermes’s published coverage. A model list, connection test or echoed reply does not prove task completion.

Hermes remains early support; image handling also depends on the generated model fields.

If something is off

Symptom	What it usually means	What to check	More help
Connection fails with a blank key	Bridge1 requires a key.	Enter your gateway key.	Endpoint setup
Wrong model or stale entries	The endpoint needs current plain ids.	Use route1 or recopy all generated models.	Model choices
Images arrive as descriptions	supports_vision may be missing.	Keep the generated vision and context fields.	Image limits

Versions

Hermes Agent requires manual setup in Bridge1 0.7.7.

Hermes Agent: limits to check

Hermes Agent has a one-turn smoke test only. Setup may change.

Review Hermes Agent and shared client limits →

Other client setup guides

Request Bridge1 early access

Hermes Agent setup questions

Is Hermes Agent tested end to end with Bridge1?

Hermes Agent has early support: a one-turn smoke test only. It has less testing than the four end-to-end clients.

Can I leave the Hermes Agent API Key field blank?

No. Bridge1 requires the gateway key, even though Hermes labels its API Key field optional.

Does Hermes Agent pass images through the custom endpoint?

Without supports_vision in the generated endpoint configuration, Hermes sends a text description instead of the image.