Skip to main content

​
What is Usage and Billing?

Every time you make an API call, it consumes quota and generates a billing record. This guide shows you how to view your usage statistics and understand how you’re being charged.
Quota is checked before each request. If you don’t have enough quota, the request will be rejected. See Limits and Prerequisites for more details.

​
How Quota Works

Model Router uses a two-phase quota system:
  1. Pre-deduction: Before your request is processed, the system checks if you have enough quota
  2. Settlement: After the request completes, you’re charged based on actual usage
Quota check and settlement behavior
Failed requests don’t consume additional quota beyond the initial pre-deduction check. You’re only charged for successful requests based on actual usage.

​
Viewing Your Total Quota and Usage

​
In the Console Sidebar

  1. Log in to the Memorylake Console
  2. Look at the quota card in the sidebar
  3. You’ll see:
    • Available quota: How much you have left
    • Used quota: How much you’ve consumed
    • Context: Whether you’re using personal or team quota

​
On the Team Quota Page

  1. Navigate to the team quota page in the console
  2. View detailed quota information for your team
  3. See breakdown by different quota sources (personal/team/mixed)
For more details about team quotas and collaboration, see the Team Collaboration documentation.

​
Viewing Individual Call Billing

​
Step 1: Open the Logs Page

  1. Go to the Logs page in the console
  2. You’ll see a list of all your API calls

​
Step 2: View Call Details

  1. Click on any call in the list
  2. A details popup will open showing:
    • Input/Output pricing: Cost per token
    • Conversion process: How the billing was calculated
    • Actual deducted quota: How much quota was used
    • Quota source: Whether it came from personal, team, or mixed quota
Pricing and billing details in call log details
Pricing and billing details in call log details

​
Understanding Billing Details

​
Pricing Information

  • Input tokens: The cost for tokens you send to the model
  • Output tokens: The cost for tokens the model generates
  • Total cost: The sum of input and output costs

​
Quota Sources

Your quota can come from:
  • Personal quota: Your individual quota
  • Team quota: Shared team quota
  • Mixed: Combination of personal and team quota

​
Settlement

  • Billing is based on actual usage, not estimates
  • Failed requests don’t consume additional quota
  • You’re only charged for what you actually use

​
Model Pricing

You can also view model pricing information on the Available Models List page in the console. This shows pricing metadata for reference, but actual billing is based on settlement results.
Pricing shown in the model list is for reference only. Your actual charges are calculated based on real usage during the settlement phase.

​
Important Notes

  1. Quota Limits: If you run out of quota, requests will be rejected. Make sure you have enough quota before making calls.
  2. Independent Quotas: Different API keys and groups have independent quotas. Using one API key doesn’t affect another.
  3. Real-time Updates: Usage statistics update in real-time. You can monitor your usage as you make calls.
  4. Billing Accuracy: You’re only charged for successful requests based on actual token usage.

​
Troubleshooting

​
”Insufficient Quota” Error

If you get this error:
  1. Check your quota in the console sidebar
  2. Make sure you have enough quota for the model you’re using
  3. Contact your administrator if you need more quota

​
Unexpected Charges

If you see unexpected charges:
  1. Check the call details in the logs
  2. Review the input/output token counts
  3. Verify the model pricing matches your expectations