Hermes Agent

Connect Hermes Agent by Nous Research to Gab AI as a custom OpenAI-compatible endpoint and chat with any Gab AI model.

Prerequisites

• A Gab AI Plus subscription
• A Gab AI API key (gab.ai → Settings → API)
• Hermes Agent installed locally — see hermes-agent.nousresearch.com/docs

Interactive Setup (Recommended)

Hermes Agent works with any OpenAI-compatible endpoint, so Gab AI can be configured as a custom provider. Run the model picker and choose the custom endpoint option:

Interactive setup

Manual Configuration

Alternatively, set Gab AI as the provider directly in ~/.hermes/config.yaml. When base_url is set, Hermes calls that endpoint directly and authenticates with the api_key value:

~/.hermes/config.yaml

Model slug case sensitivity

The model name must exactly match a slug returned by GET /v1/models — lowercase, hyphenated, no dots. "Claude Opus 4.8" or "claude opus" will return 400 invalid_model; use claude-opus-4-8.

Discover Compatible Models

Hermes' agent features work best with models that support tool/function calling. List them with:

List function-calling models

Supported Features

• OpenAI-compatible Chat Completions (/v1/chat/completions)
• Streaming responses
• Tool/function calling with tool_calls and role: tool messages
• Vision input on multimodal models (image_url with URL, base64, or file_id)

Troubleshooting

• 401 authentication_error: re-run hermes model and re-paste your key with no leading or trailing whitespace.
• 403 plus_required: the developer API requires an active Gab AI Plus subscription.
• 400 invalid_model: the model name doesn't match a real Gab AI slug. Verify against /v1/models.
• Empty or broken replies: confirm the base URL is https://gab.ai/v1 (with the /v1 suffix) and verify the endpoint with a curl request first.
• 402 credits_exhausted: your account is out of credits — check GET /v1/credits.

Multiple endpoints

If you use several OpenAI-compatible providers with Hermes, define Gab AI under the custom_providers list in config.yaml and switch during a chat with /model custom:gab-ai:arya.

hermes model
# Select "Custom endpoint (self-hosted / VLLM / etc.)", then enter:
#   API base URL: https://gab.ai/v1
#   API key:      YOUR_GAB_API_KEY
#   Model name:   arya   (or any slug from /v1/models)

model:
  default: arya
  provider: custom
  base_url: https://gab.ai/v1
  api_key: YOUR_GAB_API_KEY

Model slug case sensitivity

The model name must exactly match a slug returned by GET /v1/models — lowercase, hyphenated, no dots. "Claude Opus 4.8" or "claude opus" will return 400 invalid_model; use claude-opus-4-8.

curl https://gab.ai/v1/models \\
  -H "Authorization: Bearer YOUR_GAB_API_KEY" \\
  | jq -r '.data[] | select(.capabilities.function_calling and .capabilities.text) | .id'

Multiple endpoints

If you use several OpenAI-compatible providers with Hermes, define Gab AI under the custom_providers list in config.yaml and switch during a chat with /model custom:gab-ai:arya.

Last updated: 2026-07-28