Two ways your AI usage is measured
AI Workflow Automation tracks cost in two separate places, because there are two ways to power your AI. Which one you read depends on how the work ran.
- Local, with your own key (BYOK). Your provider bills you for tokens. The plugin’s built-in Cost Management dashboard estimates that spend so you can see it inside WordPress.
- Cloud or keyless (credits). Work metered on your connected account is measured in credits, and you read it in the Usage Center, not the Cost Management dashboard.
To understand which mode a workflow uses, see Local vs Cloud execution and BYOK keys vs keyless credits.
Local BYOK spend: the Cost Management dashboard
When you run on your own key, AI calls are billed by your provider. The Cost Management dashboard gives you an estimate of that spend inside WordPress. Open it from Settings, then Cost Management.
Time range
View your data over the last 7, 30, or 90 days, the last year, or a custom range, so you can look at both recent usage and long-term trends.
Charts
Interactive charts show your daily cost, the number of API calls, and token usage over time. Click a metric to switch views, and select a model in the chart to focus on it.
Key metrics
The top of the dashboard shows three totals for the selected period:
- Total cost: the estimated amount spent with your AI providers.
- Total API calls: how many requests were made.
- Total tokens: the volume of text processed.
Custom cost rates
Model prices change, and you may have negotiated rates. In the Model Cost Settings section you can adjust the input and output cost for any model and save. Those rates are then used for all estimates in the dashboard.
Cloud and keyless spend: credits
When a workflow runs in Cloud, or an AI node uses keyless AI, the cost is measured in credits on your connected account instead of provider tokens.
How a credit charge is built
A cloud run is charged a small base credit for the run, plus the credit cost of each node in it. Free structural nodes cost nothing, so you only pay for the steps that do real work.
Typical per-node costs
Costs are outcome-based and roughly:
- Triggers, conditions, parsers, human input, and most WordPress actions: 0 credits.
- An AI generation: 1 to 3 credits (for example, the AI node is 2, article writer 3, sentiment 1).
- A chat reply: about 3 credits.
- Sending email or calling an API or a connected app: 1 credit.
- Web scraping with Firecrawl: about 5 credits.
- Generating an image or media: about 10 credits.
- Deep web research: much higher (around 40 credits), because it runs many searches.
The builder shows an estimated-credits chip on a workflow before you run it in Cloud, and the platform reserves the real amount at run time. For current, authoritative figures, see the pricing page.
Reading your credit history
The Usage Center shows your daily consumption over the last 30 days and a running log of every credit event, grouped into usage, purchases, grants and bonuses, and refunds and adjustments. See Your account and credits for a full tour, and Plans, top-ups, and billing for adding more.
Keeping costs down
The same habits help whether you pay a provider or spend credits:
- Match the model to the task. Smaller, cheaper models handle simple jobs well; save the powerful models for hard ones.
- Write tight prompts. Shorter, clearer prompts use fewer input tokens and often give better answers.
- Set limits. Use rate limiting and maximum response lengths in your Chatbot nodes.
- Watch for spikes. A sudden jump in the chart can mean a loop misconfigured to repeat too often, or unusual traffic.
- Test before you publish. Preview a workflow to sanity-check output before it runs at volume.
Troubleshooting
Local cost data not updating? Estimates refresh daily. Check that your provider key is valid and your workflows are configured correctly.
Unexpectedly high spend? Look for spikes in the chart, check whether a Loop node is repeating more than intended, and review rate limits on your Chatbot nodes.
Credit balance looks wrong? Open the Usage Center and check the recent activity log, where each charge, grant, and adjustment is itemized.