Understanding how credits are consumed at various levels is key to monitoring usage, optimizing billing, and managing your customer activity. Our API provides all the metadata you need to track usage within your own application or business intelligence (BI) environment.
This guide explains how to implement user-level credit tracking and aggregate consumption across a customer's users or tenants.
Overview
Every stream, transcript, and document created through the API returns metadata that includes credit usage details. By storing and analyzing this data, you can calculate credit consumption for your individual users, tenants, or time periods.
You can track usage at multiple levels by:
Storing the unique interaction ID returned by the API
Using that ID to tie usage back to your internal user records
Saving all related records that include credit consumption responses (streams, transcripts, and documents)
Aggregating the
creditsConsumedfield to calculate total usageStoring and visualizing results in your own reporting or BI system
Step 1: Store the Interaction ID Returned from the API
When you create an interaction, the API response includes a unique interactionId.
Store this value in your database and associate it with the relevant user ID or customerID from your application.
Example response (simplified):
{ "interactionId": "int_abc123", "assignedUserId": "user_12345", "createdAt": "2025-10-23T10:15:30Z" }Best Practice:
Use your internal
userIdortenantIdfields to mapinteractionIdvalues to users.This mapping allows you to aggregate all downstream credit usage (e.g., transcripts, documents, or facts) associated with each interaction.
Step 2: Retrieve Usage Data
There are three responses in which Corti will provide the credits consumed within an Interaction.
Usage data is returned within the /stream endpoint once an "end" message is sent and the audio buffer has been flushed within the websocket.
Corti will return a response with a data type "usage" and the total credits consumed by the streaming session.
{
"type":"usage",
"credits":0.1
}
You can also retrieve credit consumption data from the response to a POST call that generates a transcript or a document.
POST Create Transcript
The create transcript response includes:
creditsConsumed: credits used for transcript and audio processingcreatedAt: timestamp for time-based tracking
POST Generate Document
Each generate document response includes:
creditsConsumed: credits used for document creationcreatedAt: timestamp for time-based tracking
Corti recommends retrieving and storing this data upon entity creation as this information will not be able to be retrieved down-stream.
Step 3: Aggregate and Analyze Usage
Once you have collected all relevant records, aggregate the total creditsConsumed values to calculate:
Total usage per interaction
Total usage per user (using your internal mapping)
Total usage per tenant
Usage over a specific time window
Example (simplified):
Interaction ID | User ID | Tenant ID | Date | Credits Consumed |
int_abc123 | user_12345 | tenant_A | 2025-10-01 | 120 |
int_xyz987 | user_67890 | tenant_A | 2025-10-01 | 75 |
int_qwe456 | user_98765 | tenant_B | 2025-10-02 | 300 |
Aggregate results to show total usage by user or tenant in your reporting system.
Step 4: Store and Visualize in Your BI Tool
To make this data accessible and useful:
Store all retrieved records (
interactionId,userId,tenantId,creditsConsumed, andcreatedAt) in your own database.Connect that data to a BI tool or your platform’s native reporting interface.
Use filters and visualizations to track:
Credit usage trends over time
High-usage users or customers
Billing or quota consumption reports
Example Workflow Summary
Create Interactions → Capture and store
interactionIdwith your user and customer mappingRetrieve Records → Upon ending a websocket stream or upon creation of a transcript or document retrieve Interactions, streaming usage, Transcripts and Documents.
Extract Credits → For streaming, use 'credits' and all other endpoints, use
creditsConsumedandcreatedAtAggregate Data → Sum credits by interaction, user, tenant, or date range
Store & Visualize → Save to your database or BI tool for analysis
Have a question for our team?
Click Support in the bottom-left corner of the console to submit a ticket or reach out via email at [email protected] and we'll be happy to assist you.