Manage Billing and Usage

Monitor AI for Process credit consumption and resource costs across workflows, models, guardrails, OCR models, and custom scripts. Track usage trends, set up alerts, and manage account limits to prevent service disruptions.

The Billing section in AI for Process helps you monitor expenses through detailed usage metrics and cost breakdowns. Track computational costs across workflows, models, guardrails, OCR models, and custom script deployments to make data-driven decisions about resource utilization.

Account Credits and Limits

New accounts receive the following free credits by default:

• 75 model credits — used for model inference and setting guardrails.
• 10,000 workflow runs — consumed when inferring workflows.

Once free credits are exhausted, you can no longer perform these actions. To continue using AI for Process services, contact AI for Process Support to add credits.

Accessing the Usage Page

1. Log in to AI for Process and click Settings.
2. On the left navigation pane, click Billing > Usage.

The Usage page displays the following tabs:

Tab	What it shows
Overview	High-level summary of credit and run consumption with usage trends.
Workflows	Workflow activities and operational status.
Models	Computational costs for specific models.
Guardrails	Guardrail statistics and credit consumption.
Custom Scripts	Statistics and credit usage trends for deployed custom scripts.
OCR	OCR model deployments and credit consumption.
Browser Automation	Credit consumption for browser automation pod deployments.

Use the Calendar bar on all tabs to filter usage by 24 hours, 7 days, 30 days, 90 days, or a custom date range.

Quick Usage Check

A Clock icon on the top right of the page shows a real-time usage summary. Click it to view:

• Your unique Account ID. Click the Copy icon to share with support for account debugging.
• Credits — total allocation and remaining credits for models, guardrails, and custom scripts.
• Workflow Runs — total allocation and remaining runs for workflow automation.

Click Manage billing to open the detailed Usage page.

The icon’s color reflects your remaining balance:

Color	Status	Credits remaining
Green	Good balance	More than 75%
Yellow/Orange	Moderate usage	25-75%
Red	Low credits	Less than 25%
Triangle (warning)	Credits exhausted	None

Alerts and Notifications

AI for Process sends alerts when credits are running low so you can take action before service is disrupted.

Banner Alerts

Banners appear at the top of the page in the following cases:

• Low credits or runs — A removable banner appears at 20%, 50%, 70%, 80%, and 90% credit exhaustion, encouraging you to top up or upgrade.
• Credits or runs expired — A fixed banner appears prompting you to upgrade or top up your plan.
• Free credits expired — A fixed banner appears prompting you to upgrade your plan.

Example: “Your model credits and workflow runs are low, and fine-tuning is disabled. Please contact AI for Process Support for uninterrupted service.”

Email Notifications

The system sends automated email notifications at 50%, 70%, 80%, and 90% credit exhaustion. Each email includes instructions on next steps, such as contacting support to add credits.

Example: “This is to inform you that your model credits are getting low. You cannot start a new model fine-tuning job unless you have more than 25 model credits. To ensure uninterrupted service, kindly top up your plan with more model credits or contact AI for Process Support.”

Usage Overview

The Overview tab summarizes expenses and consumption patterns across all resources:

• Runs — Workflow run usage against total allocation. For example, 45 out of 10,000 runs = 0.45% consumed.
• Credits — Credits used across models, guardrails, OCR, and custom scripts against total available credits. For example, 212 out of 300 credits = 70.72% consumed.
• Usage trend — Visual chart showing consumption over time. Use the calendar to view changes over a defined period.

A dynamic pie chart shows credit distribution by usage type (models, guardrails, OCR, and custom scripts). Each type is color-coded and labeled in the legend. Hover over the chart to see actual values.

Click the three-dots icon in the Overview tab to navigate directly to a specific usage tab.

Workflows Usage

The Workflows tab lists all workflows deployed at least once. In-development workflows are excluded.

Field	Description
Total workflows	Total number of workflows in the account.
Total runs	Total runs across all workflows.
Workflow name	Name of the workflow.
Runs	Number of times the workflow was inferred.
Owner Name	User who created the workflow.
Last active on	Date the workflow was last active.
Status	Workflow status: Deployed, Undeployed, or Deleted.

Models Usage

The Models tab lists open-source and fine-tuned models with their computational costs for storing, fine-tuning, and hosting.

Metric	Description
Total models	Total number of models in the account.
Total credits	Total credits used by all models, including all deployments.
Fine-tuning credits	Credits used for fine-tuning, based on training data size, model complexity, training epochs, hardware type, and other parameters.
Hosting credits	Credits for deploying and hosting models on GPUs. Each GPU instance is a “replica.” For example, one A100 GPU = one replica; two A100 GPUs = two replicas, charged at double the rate.

Each model record shows: Model name, Type (Fine-tuned or Base), Credits used, and Last deployed on.

Deployment-level Information

Click a model record to view all its deployments. The following information is displayed:

• Name — Name given to the deployment.
• Type — Type of deployment.
• Credits used — Credits consumed by the deployment.
• Last updated on — Date of the last deployment.
• Status — Deployed, Undeployed, or Deleted.

Guardrails Usage

The Guardrails tab lists all guardrails and their hardware-based charges. Guardrail usage is deducted from Model credits shown in the Overview tab.

Field	Description
Total guardrails	Total guardrails used in the account.
Total credits	Total credits used by all guardrails.
Time	Total duration all guardrails were deployed.
Name	Name of the guardrail or scanner.
Credits	Credits used to deploy the guardrail.
Time	Duration the guardrail was deployed.
Last deployed on	Date the guardrail was last deployed.
Status	Deployed or Undeployed.

Credit behavior when credits are low:

• New guardrail deployments are disabled when credits fall below the lowest threshold.
• An active deployment continues and proceeds into negative credits if the low credit limit is reached mid-deployment.
• Once the negative credit limit is crossed, the deployment stops with the message: “You’ve used all your available credits. Please add more credits to your account to continue.”
• Negative credits are adjusted in the next billing cycle.
• When no credits exist, further deployments are disabled and all active deployments stop when the negative credit limit is reached.

Detailed Guardrail View

Click a row to open a right-side panel showing:

• The guardrail name.
• Hardware — CPU type, number of CPUs, usage duration, and credits used.
• Total — Total credits used for the hardware.

Email Notifications

Billing emails for guardrails cover credit usage, negative threshold breaches, and potential deployment interruptions.

Custom Scripts Usage

The Custom Scripts tab lists deployed custom scripts with their credit consumption. Total Scripts and Hosting Credits are shown as key metrics.

Field	Description
Total Scripts	Total scripts added and managed in the account.
Hosting Credits	Total credits used by all scripts.
Script name	Name of the script.
Language (Version)	Language and version of the script.
Credits used	Credits consumed for hardware and resources.
Last active on	Latest date the script was actively used.
Status	Deployment status. See script deployment statuses.

Credits are deducted based on the hardware profile and resource allocation table.

Credit behavior when credits are low:

• New deployments are disabled when credits are insufficient.
• An active deployment continues and proceeds into negative credits if the low credit limit is reached mid-deployment.
• Once the negative credit limit is crossed, the deployment stops with the message: “You’ve used all your available credits. Please add more credits to your account to continue.”
• Negative credits are adjusted in the next billing cycle.
• When no credits exist: projects can only be imported as drafts, deployments are disabled, and active deployments stop when the negative credit limit is crossed.

Detailed Script View

Click a row to open a right-side panel showing:

• Hosting infrastructure — Hardware configurations (vCPUs and memory).
• Hosting time in hours and minutes.
• Credits consumed by components. See pricing.
• Total — Aggregate of all credit components.

Email Notifications

Billing emails for custom scripts cover credit usage, negative credit threshold breaches, and deployment or undeployment events. Deployment notifications also include API endpoint access details.

OCR Usage

The OCR tab provides a unified view of all OCR-related credit usage. Use it to:

• Review OCR model deployments.
• See credits consumed per deployment.
• Check deployment status and usage history.

Detailed OCR View

Click a row to open a right-side panel showing:

• Hosting infrastructure — Compute resources allocated to the deployment.
• Hosting time — Total active time in hours and minutes.
• Credits consumed — OCR credits used by the deployment.
• Total — Aggregated credit usage across all contributing components.

Browser Automation Usage

The Browser Automation tab provides a unified view of browser automation credit usage. Use it to:

• Review pod deployments for browser automation scripts.
• See deployment-level credit consumption.
• Check deployment status and usage history.

Detailed Browser Automation View

Click a row to open a right-side panel showing:

• Hosting infrastructure — Compute resources or allocated hardware used by the deployment.
• Hosting time — Total active time in hours and minutes.
• Credits consumed — Browser automation credits used by the deployment.
• Total — Aggregated credit usage across all contributing components.

Best Practices

• Regularly monitor usage trends to optimize resource allocation.
• Review workflows, models, guardrails, OCR, and custom scripts to identify cost-saving opportunities.
• Check the hardware profile and credits consumption chart before deploying custom scripts.
• Track guardrail deployment duration to manage hardware costs effectively.

For help with adding credits or managing your account, contact Support.