Quickstart

Go from zero to a working API request in a few minutes.

Prerequisites

You need a Vibetracker account and an API key. Create your API key and store it as an environment variable:

bash

export VIBETRACKER_API_KEY="vbt_ak_your_prefix_your_secret"

The plaintext key is only shown once. Copy it before closing the dialog.

Send a request

The minimum payload is model and score.

curl -X POST 'https://vibetracker.app/api/v1/opinions' \
  -H 'Authorization: Bearer $VIBETRACKER_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "gpt-4o",
    "score": 1
  }'

Add context

Include optional fields like useCase, interface, toolId, and comment for richer sentiment data. When sending tool context, interface becomes required.

curl -X POST 'https://vibetracker.app/api/v1/opinions' \
  -H 'Authorization: Bearer $VIBETRACKER_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "gpt-4o",
    "score": -1,
    "updateOptionalContext": true,
    "useCase": "coding",
    "interface": "api",
    "toolId": "openai-api",
    "comment": "Regression in tool calls after the latest deploy."
  }'

Response

A successful request returns the stored opinion. Check createdNewOpinion to know if the call created a new opinion or updated an active one.

json

{
  "opinionId": "jh7m9w7r9m2w0n9w7d3h6n3w3n7c0sd0",
  "createdNewOpinion": true,
  "model": {
    "displayName": "GPT-4o",
    "fullSlug": "openai/gpt-4o"
  },
  "score": 1,
  "useCase": null,
  "interface": null,
  "toolId": null,
  "toolNameOther": null,
  "comment": null,
  "moderationStatus": "visible",
  "submittedAt": 1773686400000,
  "cooldownEndsAt": 1773688200000,
  "currentSentiment": {
    "model": {
      "displayName": "GPT-4o",
      "fullSlug": "openai/gpt-4o",
      "providerName": "OpenAI",
      "providerSlug": "openai",
      "status": "active"
    },
    "sentiment": {
      "label": "Positive",
      "netSentiment": 0.5,
      "totalSubmissions": 8,
      "recentDelta": 0.12,
      "baselineSentiment": 0.38,
      "positiveCount": 5,
      "neutralCount": 2,
      "negativeCount": 1,
      "positiveShare": 0.63,
      "neutralShare": 0.25,
      "negativeShare": 0.13
    },
    "sparkline": [
      { "day": 1773619200000, "avg": 0.5, "count": 8 }
    ]
  }
}

Common mistakes

Missing the Bearer prefix on the Authorization header
Sending a score that is not -1, 0, or 1
Using a short model slug that matches multiple providers. In that case, send the canonical provider/model slug instead.
Passing a toolId that does not belong to the selected interface
Including links in comment or toolNameOther

For the full field-by-field reference, see POST /api/v1/opinions.