Skip to main content

​
Background

Until the Sunny Seal release, the platform included a built-in Analytics page.
This page allowed customers to track metrics such as:
  • request volume,
  • user satisfaction,
  • overall usage statistics.
Over time, however, we realized that every customer had very different reporting needs. A single, one-size-fits-all dashboard could not satisfy everyone. That’s why we decided to deprecate the Analytics page starting with Sunny Seal.

​
New Strategy: Flexibility through the API

Instead of enforcing a generic interface, we shifted to a more flexible approach.
Now, usage tracking is enabled through a series of API endpoints that expose raw data. This choice has two main benefits:
  1. Total flexibility – each customer can build dashboards and reports tailored to their needs.
  2. Easy integration – data can be plugged directly into existing systems (BI tools, CRM, internal monitoring, etc.).

​
Available Endpoints

The following endpoints are available for reporting:

​
How to Use the Endpoints

​
Authentication

Each request requires an API token passed as a Bearer token in the header: 
Authorization: Bearer <YOUR_API_TOKEN>

​
Common Parameters

  • start_date (required): start date in YYYY-MM-DD format.
  • end_date (optional): end date, defaults to today.
  • Depending on the endpoint, additional filters may be available (e.g., user ID, model ID).

​
Example: Get Chat Sessions Report

​
cURL

curl --request GET \
  --url "https://paradigm-preprod.lighton.ai/api/v2/reporting/chat-sessions/?start_date=2023-12-01&end_date=2023-12-31" \
  --header "Authorization: Bearer <YOUR_API_TOKEN>"

​
Python

import requests

url = "https://paradigm-preprod.lighton.ai/api/v2/reporting/chat-sessions/"
params = {"start_date": "2023-12-01", "end_date": "2023-12-31"}
headers = {"Authorization": "Bearer YOUR_API_TOKEN"}

response = requests.get(url, headers=headers, params=params)
data = response.json()

print(data["reporting"][0]["dates"][0]["sessions"])

​
Example JSON Response

{
  "metadata": {
    "sessions_count": {"description": "Total number of sessions", "unit": "integer"},
    "average_session_length": {"description": "Average session length", "unit": "seconds"}
  },
  "reporting": [
    {
      "company": {"name": "ACME Corp", "id": 123},
      "dates": [
        {
          "date": "2023-12-25",
          "sessions": {
            "sessions_count": 45,
            "unique_models_count": 3,
            "average_session_length": 120
          },
          "chats": {
            "user_queries_count": 200,
            "model_responses_count": 190,
            "last_model_response_date": "2023-12-25T22:30:00Z"
          }
        }
      ]
    }
  ]
}

​
Example Use Cases

  • BI dashboards: integrate with PowerBI, Looker, or Tableau to track adoption.
  • Internal alerts: trigger notifications (e.g., threshold exceeded).
  • Product insights: monitor which tools or documents are most used.

​
FAQ

​
What filters are available?

All endpoints support date filters, and some accept additional filters (e.g., user ID, document ID). See the API reference for details.

​
How often can I call the API?

You can query as often as needed, but we recommend aggregating data daily or weekly to reduce load.

​
Do you still provide a built-in reporting interface?

No. Our strategy is API-only for maximum flexibility and scalability.

​
Can I combine multiple endpoints?

Yes. For example:
  • Sessions + Feedbacks = adoption & satisfaction tracking.
  • Documents + Tools = resource usage monitoring.
⌘I