Always getting the error "Fail to save evaluation" on evals every time

Thanks for reporting the issue! To help diagnose the "Fail to save evaluation" error, could you please let me know if you see any specific error messages in your browser's console or in the API logs when you try to save the evaluation? This information will help pinpoint the cause. (You can open the console with

F12

in most browsers and check the "Console" tab for errors.) [More troubleshooting tips here](https://docs.vapi.ai/issue-reporting) Sources: - [How to Report Issues Effectively](https://docs.vapi.ai/issue-reporting) - [Debugging voice agents](https://docs.vapi.ai/debugging)

Im getting a problem similar to this

Hi jm, there could be a few potential issues to check related to your eval configuration: 1. Make sure your evaluation model settings are correct: - Choose from supported models: GPT-4.1, o1-mini, o3 (OpenAI), Claude (Anthropic), Gemini (Google), or custom models - Set appropriate token limits (recommended 50-10000 tokens) - Use temperature between 0-0.3 for LLM-as-a-judge to reduce hallucinations 2. Ensure proper evaluation setup: - The judge model must respond with exactly "pass" or "fail" - Design evaluation prompts to ensure clear, deterministic responses - Check that your `EvalRun`\] configuration is complete 3. Verify your evaluation target is properly specified: - For assistant targets: Use `EvalRunTargetAssistant`\] - For squad targets: Use `EvalRunTargetSquad`\] 4. Monitor the evaluation status: - Check for proper status transitions (queued -> running -> ended) - Look for specific

endedReason

values that might indicate why it's failing[1](https://docs.vapi.ai/changelog/2025/9/28) Without more specific error details, I recommend checking these common configuration points. If the issue persists, you may want to check the evaluation response format and ensure it strictly adheres to the pass/fail requirement. If this doesn’t resolve the issue, please let us know and we’ll be happy to step in and assist further.