How to fix “missing required parameter” errors from OpenAI node in n8n?

The fastest and most reliable way to fix “missing required parameter” errors in the OpenAI node is to check which model you selected and then explicitly fill in every field that model requires. In practice, these errors happen when the model expects inputs like messages, input, model, or response\_format, but one of them is empty, coming through undefined from a previous node, or hidden behind a conditional UI section that n8n didn’t send to the API. Fixing it usually means: pick the right operation in the node, make sure expressions actually resolve to real values, and match your fields to the model’s expected schema.

 

Why this error actually happens in n8n

 

OpenAI changed their API models and schemas several times. The n8n node mirrors those requirements. When the OpenAI node runs, n8n sends a JSON body to the API. If any required field is null, empty string, or missing, OpenAI returns the “missing required parameter” error.

The most common real‑world causes:

• You selected a model like gpt-4o-mini, gpt-4o, or o1 but the messages field is empty or invalid.
• You selected the Embeddings operation but left input empty.
• You’re using expressions such as {{$json.text}} that resolve to undefined because the previous node didn’t output the field you expected.
• You switched models and the UI changed, hiding a field you previously filled in, so the request no longer matches the model schema.
• You copied old OpenAI node settings from before the API update and some field names no longer line up.

 

How to fix it reliably

 

Here is the approach I use in production workflows when debugging these errors:

• Check the operation in the node. Make sure it matches the type of request you need (Chat, Embeddings, Images, etc.). Different operations have different required parameters.
• Check the model. The n8n node only shows fields valid for that model. For example, all chat models require a valid messages array. Embedding models require input. If you switch models, recheck every field.
• Test expressions by hovering over them. n8n will show you what value the expression resolves to. If it shows undefined or empty, that is your cause. Fix the upstream node so the field exists.
• Hard‑code a simple value temporarily to verify the node works. For example:
  Hello from n8n inside the messages field. If this works, your dynamic data was the issue.
• Inspect the raw input using the node's “Show JSON” to confirm the request body contains every property. What you see in the UI is not always what goes into the JSON.
• Ensure you aren’t sending empty arrays. For example: [ ] is still considered “missing” for OpenAI.
• If using message arrays, always include role and content: { "role": "user", "content": "Hello" }

 

Example of a working Chat request

 

Below is a minimal valid message setup inside an n8n Function node that you can pass into the OpenAI node to confirm correctness:

// This produces a valid messages array for an OpenAI Chat model
return [
  {
    json: {
      messages: [
        { role: "user", content: "Say hi!" }
      ]
    }
  }
]

Then in your OpenAI node, set the messages field to:

{{$json.messages}}

If that works, then any missing‑parameter issues come from your dynamic message building.

 

One more common production pitfall

 

If you’re merging items from multiple nodes, you may end up with multiple items where only the first contains the field you want. The OpenAI node processes items one by one. If item #2 is missing messages, it will fail even if item #1 works. The fix is usually to wrap your data using a Function node that consolidates everything into one item with all required fields.

 

In summary: validate the fields required by the specific model, verify every expression resolves to a real value, and ensure the node receives exactly one complete JSON item with all required parameters. That's how you fix “missing required parameter” errors in the OpenAI node in a reliable, production‑safe way.