Handling tool result when there’s no real output (form filling use case) #8182
              
                Unanswered
              
          
                  
                    
                      TomerFisher
                    
                  
                
                  asked this question in
                Help
              
            Replies: 1 comment
-
| You can either return a dummy result that makes sense for the LLM or use data parts instead of remove that particular tool call/response. | 
Beta Was this translation helpful? Give feedback.
                  
                    0 replies
                  
                
            
  
    Sign up for free
    to join this conversation on GitHub.
    Already have an account?
    Sign in to comment
  
        
    
Uh oh!
There was an error while loading. Please reload this page.
-
My use case is a chat assistant that helps users fill out a form.
• I created a tool with an input schema that matches the form fields.
• On the client side, I take the tool’s input and use it to populate the form.
The issue: The SDK requires me to return a result (output) from the tool, but in my case there isn’t really an output — the important part is just passing the input values to the form.
What’s the correct way to handle a tool like this in the Vercel AI SDK? Should I just return a dummy output, echo the inputs back, or is there a recommended pattern for “input-only” tools?
Beta Was this translation helpful? Give feedback.
All reactions