API429 in n8n: Header Auth and a manual Chat Completions workflow

Prepare a single-item manual test

Start in a separate test workflow with one input item and a short nonconfidential prompt. Obtain your API429 client key and compare gemini-3.8-flash with GET /v1/models for that key. The public model catalog is useful for discovery, but it does not establish access under every credential. Check the exact spelling rather than choosing a similar display name.

The example uses Manual Trigger followed by HTTP Request typeVersion 4.3. It contains no schedule, webhook or saved credential and has active set to false. Reading the JSON or validating its structure is not a native n8n import test. Native client execution and a paid model request were not performed for this guide. Importing the file should be followed by inspection in your own installed version.

Create and attach Header Auth securely

In n8n, create a Header Auth credential. Set the header name to Authorization and enter Bearer followed by your API429 key as the protected value. Select this credential on the HTTP Request node under Generic Credential Type and Header Auth. Do not put the real key in the URL, JSON body, node notes or an exported workflow.

The downloaded node uses generic httpHeaderAuth. It is not an OpenAI credential and should not be substituted with one in this template. n8n also exposes a Base URL field for OpenAI credentials, which can be useful for separately supported node operations. That separate option does not prove that every OpenAI node method or AI Agent feature matches this specific Chat Completions contract.

Inspect the request before executing

The HTTP node uses POST https://gateway.api429.com/v1/chat/completions. Send Body is enabled, content type is JSON and the body is specified as JSON. Its literal fields are model, messages, max_tokens and stream. Keep max_tokens as the number 256 and stream as the boolean false. The single user message asks for a one-sentence API explanation; no expression or incoming-item mapping is required.

The value 256 bounds this example output and does not describe the maximum output limit of the model. The workflow includes no custom header injection, extra nodes or automatic retry settings. Review these properties after import. Clicking Execute workflow submits a model request with the attached credential and can incur a charge even though the saved workflow is inactive.

Read and map the text result explicitly

After a deliberate manual run, inspect the HTTP response JSON. A normal text Chat Completions response places the answer in choices[0].message.content. Confirm that the field exists and has the expected type before using it. The two-node example leaves the result visible in the HTTP node; it does not hide errors or automatically forward output elsewhere.

When you later connect another node, select the response field explicitly using the actual output shown in your run. Keep manual mapping and validation in your own workflow. Do not assume the result has a Gemini-native candidates structure because the selected model belongs to Google. The method here uses the OpenAI-compatible Chat Completions envelope. Image, tool and structured-output operations need their own contract checks.

Diagnose failures and count real usage

For authentication errors, inspect the selected credential and confirm the Authorization header format without exposing its value. For a model error, compare the exact ID against the key-specific catalog. For body errors, verify JSON types: messages is an array, max_tokens is numeric and stream is boolean. Inspect a redacted error rather than turning failures into successful downstream data.

A workflow processing several items may make several requests. Begin with one item, then measure actual API429 usage before increasing volume. Retries may also cost money; avoid unbounded repetition. n8n hosting or subscription costs are separate from model charges. A successful text response establishes only that request, and does not validate AI Agent, tool calling or every specialized OpenAI node operation.

Inactive manual workflow: import JSON

{
  "name": "API429 — manual Chat Completions",
  "nodes": [
    {
      "parameters": {},
      "id": "714a40c3-ec65-44c9-b90a-270ad0aa0c9c",
      "name": "Manual Trigger",
      "type": "n8n-nodes-base.manualTrigger",
      "typeVersion": 1,
      "position": [
        0,
        0
      ]
    },
    {
      "parameters": {
        "method": "POST",
        "url": "https://gateway.api429.com/v1/chat/completions",
        "authentication": "genericCredentialType",
        "genericAuthType": "httpHeaderAuth",
        "sendBody": true,
        "contentType": "json",
        "specifyBody": "json",
        "jsonBody": "{\n  \"model\": \"gemini-3.8-flash\",\n  \"messages\": [\n    {\n      \"role\": \"user\",\n      \"content\": \"Answer in one sentence: what is an API?\"\n    }\n  ],\n  \"max_tokens\": 256,\n  \"stream\": false\n}",
        "options": {}
      },
      "id": "286f236b-773d-4793-84d5-2df08744d9eb",
      "name": "API429 Chat",
      "type": "n8n-nodes-base.httpRequest",
      "typeVersion": 4.3,
      "position": [
        280,
        0
      ]
    }
  ],
  "connections": {
    "Manual Trigger": {
      "main": [
        [
          {
            "node": "API429 Chat",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  },
  "active": false,
  "settings": {
    "executionOrder": "v1"
  }
}

Payment and a $10 test

• Russian bank card. Top up your balance in the account.
• Invoice for companies. Bank transfer with closing documents.
• USDT and TON. Top up in crypto.

$10 test credit. Subscribe to our Telegram channel through the bot and get $10 on your balance for the first requests. Get $10 in the bot