The Capture Feedback API allows you to collect and store user feedback on AI interactions. This feedback can be used to improve the quality of AI responses and track user satisfaction.
Available Endpoints
Set Conversation Score
Set a score for an entire conversation based on the conversation id, providing feedback on the overall quality of the interaction.
PATCH /conversations/{conversationId}/scoreSet Messages Score
Set a score for specific messages within a conversation based on the conversation id and message id.
PATCH /conversations/{conversationId}/messages/{messageId}/scoreSet Prompt Run Score
Set a score for a specific prompt run based on the prompt run id.
PATCH /prompt-runs/{promptRunId}/scoreRequest Format
The request format includes:
- A numerical score (e.g., 1-5)
- Optional text feedback
- Optional tags or categories for the feedback
Response Format
The response confirms that the feedback has been recorded successfully.
Use Cases
- Collecting user satisfaction metrics
- Identifying problematic AI responses
- Training and fine-tuning AI models based on user feedback
- Creating reports on AI performance
PATCH / conversations / { conversationId } / score;
PATCH / conversations / { conversationId } / messages / { messageId } / score;
PATCH / prompt - runs / { promptRunId } / score;