How can I fix the empty object properties error in the Google Gemini Chat Model node in my n8n workflow?

The error message suggests that one of the tool function declarations configured in your Google Gemini Chat Model node has an object-type parameter with an empty properties field. In other words, if you’re using tools, the Gemini API expects that for any parameter defined as type OBJECT, you provide a valid schema via the properties. Here are a few steps to help troubleshoot and resolve the issue:

1. Check the Tools Configuration

If you have defined a tool (or function) in your AI Agent node, review its configuration. In particular, ensure that for any tool with a parameter defined as an object, you have filled out the "properties" section with a valid schema. For example, instead of leaving:

JSONCopied
"parameters": {
  "type": "object",
  "properties": {}
}

You should define at least one property:

JSONCopied
"parameters": {
  "type": "object",
  "properties": {
    "location": {
      "type": "string",
      "description": "The city and state, e.g. Chicago, IL"
    }
  },
  "required": ["location"]
}

2. Verify Your Use Case

Decide if you really need to use the tool functionality. If your workflow doesn’t require making function calls via the Gemini API, you can remove or disable the tool configuration in the Google Gemini Chat Model node. This will bypass any function declaration issues.

3. Cross-Reference Documentation

Review the Google Gemini Chat Model node documentation for details on how to set up the node parameters and tools properly. This can help ensure that your tool function declarations adhere to the expected schema.

Following these steps should help you resolve the error and allow your workflow to run smoothly. Let me know if you need any further clarification!

Below are some steps to help resolve the error:

What's Happening

The error indicates that the request you’re sending to Google Gemini is failing validation. In particular, the error message:

  GenerateContentRequest.tools[0].function_declarations[0].parameters.properties: should be non-empty for OBJECT type

means that a function declaration in your request was provided with an empty or missing properties object for its parameters.

How to Fix It

1. Check Your Node Configuration:
   Verify the settings in your AI Agent node (Google Gemini Chat Model). If you’re not using function call support, make sure you haven’t added any unnecessary function declarations.

   • If you are not using functions: Remove or disable any function-related configuration options.
   • If you are using functions: Make sure that for every function declaration, the parameters object is populated with at least one valid property. For example, instead of:

     JSONCopied
     {
       "name": "myFunction",
       "parameters": {
         "properties": {}
       }
     }

     provide valid properties:

     JSONCopied
     {
       "name": "myFunction",
       "parameters": {
         "properties": {
           "arg1": {
             "type": "string",
             "description": "A valid argument description."
           }
         }
       }
     }

2. Review the Documentation:
   Double-check the documentation for the Google Gemini Chat Model node in n8n. Documentation may include details or examples on how to correctly set up function declarations if you need them. If not using them, it might be best to leave this section out.

3. Test Your Changes:
   After updating the node configuration, test your workflow again. If the error persists, consider temporarily switching the model (if your account supports others) to see if configuration differences help isolate the issue.

Summary

• The error is due to an empty parameters properties object in a function declaration.
• Remove function call declarations if they aren’t needed.
• If you intend to use them, fill in the required properties.
• Validate your changes by running the workflow.

Following these steps should resolve the 400 Bad Request error stemming from the Gemini API.

Based on the error message, it appears that one of the function declarations you’re sending as part of your request to the Gemini API isn’t fully configured. In this case, the API expects that any parameter defined as an object must include at least one property (i.e. a non‑empty "properties" field).

Here are some steps to resolve the issue:

1. Check the AI Agent (Google Gemini Chat Model) Node Configuration:
   In your node settings, look for any tool or function declaration configuration. If you’ve set up a function that accepts parameters defined as an object, ensure that the object’s schema includes at least one property. For example, if you’re using function call capabilities, you need a valid JSON schema such as:

   JSONCopied
   {
     "type": "object",
     "properties": {
       "exampleField": {
         "type": "string",
         "description": "A description for your field"
       }
     },
     "required": ["exampleField"]
   }

2. Remove or Disable Unused Declarations:
   If you do not intend to send any function declarations (or specific tool configurations) with your request, remove or disable that configuration in your node. This way, the API won’t try to validate an empty object.

3. Confirm the Intended Use of Function Declarations:
   Review the documentation for the Gemini Chat Model node in n8n and ensure you’re aligning with how the node expects declarations to be structured. Sometimes the node automatically builds parts of the request, so double-check if there’s an option to manually override function declarations.

4. Test with a Minimal Setup:
   To isolate the issue, try running the workflow with a simplified configuration where you remove optional fields or function declarations. If the error goes away, then gradually add back your custom logic ensuring that any object-type parameter includes non-empty properties.

Following these steps should allow you to either provide the required JSON schema for your function declaration or remove the unnecessary configuration so that the Gemini API request meets its requirements.