n8n

n8n is one of the easiest harnesses to connect, because its OpenAI and Anthropic credentials expose a Base URL field.

See Connect your harness for your base URL, key, and model IDs.

1. Create or edit an OpenAI credential.
2. Set Base URL to https://app.tappass.ai/v1.
3. Set API Key to your TapPass agent API key.
4. In the OpenAI Chat Model node, set the model to a prefixed ID, for example anthropic/claude-sonnet-4-5-20250929.

The AI Agent node’s tool-calling path then works as if it were talking to the provider directly.

If a node does not expose a base URL, use an HTTP Request node:

• Method: POST
• URL: https://app.tappass.ai/v1/chat/completions (or /v1/messages)
• Header: Authorization: Bearer <your TapPass key>
• Body: a standard OpenAI or Anthropic request with a prefixed model ID.

n8n’s newer Agent Builder reduces an OpenAI credential to only { apiKey, baseURL }, silently dropping custom header fields and stripping the provider prefix from the model ID (n8n issue #39227). That breaks the two things TapPass relies on: prefixed model IDs and routing headers. Until it is fixed, use the classic LLM nodes or the HTTP Request node for gateway routing.

Run the workflow once, then confirm a governed event for your agent in the TapPass console.