Pro reporting · Cost and Token Analytics
Understand every successful GPT request without exposing the content behind it.
Follow token usage, estimated USD costs, generated articles and model activity across your publishing network. Filter the read-only Analytics view by period, website, GPT model or source and turn usage metadata into clearer operational decisions.
Read-only reporting · Content-free usage metadata · Filtered CSV and JSON exports
One view across the workflow
See where GPT tokens and estimated costs are being used
Open Analytics from the SitePublisher AI Dashboard to inspect six connected areas of usage. The view is built from recorded metadata and does not need to read the generated articles themselves.
Costs and daily trend
Review total and average estimated cost in USD, together with separate input and output costs and a daily cost chart for the selected period.
Detailed token totals
Compare input, cached input, output and reasoning tokens. See total and average usage plus the percentage of input served from cache.
Website comparison
See requests, generated articles, tokens, estimated cost and last activity for each registered website. Websites with no activity remain visible.
Article generation
Follow successful article-generator records, average token usage, total estimated cost and the latest generation time without loading article text.
Source and request activity
Compare successful requests by workflow source, including attempts, retries, average duration, tokens and estimated cost.
GPT model usage
Understand which supported models were used, their request share, token and reasoning totals and their calculated contribution to cost.
Focused snapshots
Filter the data you need, then export the same view
Analytics opens with the last 30 days. Change the scope to seven, 90 or 365 days, or include all recorded usage. Add a website, model or source filter whenever you need a narrower answer.
Tracked costs use recorded token usage and the pricing catalog included with the plugin. They are useful operational estimates, but the customer’s OpenAI billing account remains the definitive invoice.
Open Analytics from the Dashboard
The Pro dashboard provides an Open Analytics button. The reporting view is kept inside the main SitePublisher AI workspace instead of adding another permanent menu item.
Select a reporting period
Choose the last 7, 30, 90 or 365 days, or All recorded usage. Every section uses the same active period.
Narrow the snapshot
Filter by registered website, supported GPT model or generation source. Filters can be combined for more precise comparisons.
Refresh the read-only view
Generate a new snapshot from the stored usage records. Refreshing does not send an AI request or alter any WordPress content.
Export JSON or CSV
Download the aggregate data behind the active filters. Export files are generated in memory when requested and are not stored by SitePublisher AI.
Basic visibility in Free, network analysis in Pro
See individual usage first, then analyse the complete publishing operation
Free keeps the essential cost signal close to content creation. Pro turns the accumulated usage metadata into a filterable, exportable Analytics view.
Immediate generation metrics
Review the token total and estimated cost returned with a completed generation on the current website.
- Tokens and estimated cost beside the generated result
- General requests, token and tracked-cost totals on the Dashboard
- Your own OpenAI API account remains the billing source
- No dedicated multi-website Analytics view
Complete read-only Analytics
Explore usage across websites, time periods, models and generation sources from one central reporting view.
- Cost totals and daily trend
- Detailed input, cached, output and reasoning tokens
- Website, article, API source and GPT model comparisons
- Filtered JSON and CSV exports
Content-free by design
Analyse usage without reopening private content
The Analytics snapshot is deliberately limited to operational metadata. It reports what was used, where and when—without using the prompt or generated output as reporting data.
No sensitive content
- No prompts are loaded into Analytics
- No generated article output is loaded
- No OpenAI API key is loaded or exported
Read-only reporting
- No AI call is made to build a snapshot
- No posts, pages or settings are changed
- Aggregate exports are not stored automatically
Successful GPT requests
- Only successful GPT requests create usage records
- Failed requests are not included in current totals
- Image API costs remain visible in OpenAI billing
Better operational decisions
Compare before you change the workflow
Use website totals to identify where activity is concentrated, model rows to compare request share and cost, source data to spot slower or retry-heavy workflows and cached-token figures to understand efficiency. Analytics does not make automatic changes; it gives you the evidence needed to adjust models, prompts or publishing volume deliberately.
Frequently asked questions
Cost and Token Analytics
Practical answers about access, estimates, data scope, image costs and exports.
Is the full Analytics view included in Free?
No. Free shows basic token and estimated-cost information with generations and general AI totals on the Dashboard. The filterable network Analytics view is a Pro feature.
Are the displayed costs the final OpenAI invoice?
No. They are USD estimates calculated from recorded token usage and the plugin’s model-pricing catalog. Always use the customer’s OpenAI billing account as the definitive billing record.
Are AI image costs included in these totals?
The current Analytics view tracks successful GPT text requests. Image-generation charges are separate OpenAI Image API usage and should be checked in the connected OpenAI account.
What information is stored for Analytics?
Usage records contain operational metadata such as response ID, model, source, website reference, token counts, cost breakdown, attempts, duration and timestamp. Prompts, generated output and API keys are not stored in the record.
Are failed GPT requests included?
No. Only successful GPT requests create usage records in the current version. Attempts and retries that led to a successful request are retained with that record.
Can I export the filtered results?
Yes. Pro can export aggregate JSON or CSV data for the active period and filters. The report is created on demand in memory and is not retained automatically.
Make every GPT request easier to understand and manage.
Choose Pro for content-free network Analytics, clear filters and on-demand aggregate exports.