Feedback Tracking

Use feedback tracking for:

• user’s reactions to your chatbot’s responses directly on the frontend.
• score LLM outputs directly yourself

You can then use this to filter llm calls and fine-tune your own models based on the data.

Feedback tracking can be done in the backend or directly on the frontend if it’s easier for you.

const msgId = thread.trackMessage({
  role: 'assistant',
  content: 'Hello! How can I help you?',
})

lunary.trackFeedback(msgId, { thumb: 'up' })

const msgId = thread.trackMessage({
  role: 'assistant',
  content: 'Hello! How can I help you?',
})

lunary.trackFeedback(msgId, { thumb: 'up' })

msg_id = thread.track_message({
  "role": "assistant",
  "content": "Hello! How can I help you?"
})

lunary.track_feedback(msg_id, { "thumb": "up" })

​

Example of a React Feedback component

You can also use the feedback method to track user reactions to your chatbot’s responses directly on the frontend.

<Button onClick={() => lunary.trackFeedback(message.id, { thumb: "up" })}>
  👍
</Button>

The trackFeedback method takes two arguments:

• runId: the ID of the message or run you want to track the feedback on.
• feedback: an object containing the feedback data. You can use any key/value pair you want.

​

Feedback data

You can send any feedback data you want, as long as it’s a valid JSON object.

We recommend using the following keys to ensure that data is displayed correctly in the dashboard.

​

Removing feedback

To remove feedback, simply pass null as the feedback data.

lunary.trackFeedback(message.id, { thumb: null })