When your Promptitude tools aren't working as expected, it can disrupt your entire workflow. This article walks you through common issues and practical solutions to get your tools functioning properly again.
By following these troubleshooting steps, you'll be able to identify and resolve most tool-related problems without needing technical support.
Common Issues and Solutions
1️⃣ Function Definition Problems
One of the most common reasons tools fail is improper function definition. Here's how to fix it:
Check function purpose and parameters: Ensure you've provided clear descriptions of what each function does and what inputs it expects.
Review parameter descriptions: Make sure each parameter has precise descriptions that explain exactly what information is needed.
Use correct naming: When referencing a tool in your prompt, always use the exact name of the tool, not the label.
2️⃣ Formatting Issues
Incorrect formatting can prevent your tools from working properly:
Verify syntax: Make sure your function calls follow the correct format and structure.
Check for typos: Even small errors in spelling or punctuation can cause tools to fail.
Review JSON structure: If your tool uses JSON, ensure all brackets, commas, and quotation marks are properly placed.
3️⃣ Tool Selection and Relevance
Having too many tools or irrelevant tools can confuse the model:
Limit your tools: Only include tools that are directly relevant to your current task.
Remove unnecessary functions: Too many options can overwhelm the model and reduce answer quality.
Focus on task-specific tools: Use tools designed specifically for your current needs.
4️⃣ System Instructions and Context
Unclear instructions can lead to tool failures:
Provide clear guidance: Explicitly state when and how tools should be used.
Add context: Give the model enough background information to understand when tool use is appropriate.
Set proper expectations: Clearly define what you want the tool to accomplish.
5️⃣ Model Selection Issues
Different models handle tools differently:
Use appropriate models: For complex tools, use the latest models.
Consider model limitations: Simpler models like Claude Haiku may struggle with complex tools.
Check model compatibility: Ensure your chosen model supports the tools you're trying to use.
Learn all the details about function calling and best practices within Promptitude at: ➔ Function Calling in Promptitude: What You Need to Know
📌 Advanced Troubleshooting Steps
If the basic solutions don't resolve your issue, try these more advanced approaches:
Test with simplified inputs: Reduce complexity to isolate the problem.
Check API connections: Verify that your endpoint configuration is correct.
Review authorization: Ensure your Bearer token is valid and properly formatted.
Examine response paths: Confirm you're correctly specifying which part of the response to use.
Try parallel tool use: If using multiple tools, check if disabling parallel tool use helps.
🔔 When to Seek Help
If you've tried all the solutions above and still face issues:
Check documentation: Review the Promptitude help center for specific guidance.
Look for updates: Make sure you're using the latest version of the platform.
Contact support: Reach out to the Promptitude team with details about your issue.
Remember, most tool issues can be resolved by carefully reviewing your configurations and making small adjustments. With patience and systematic troubleshooting, you'll have your tools working efficiently again in no time.