Tracker MCP tools

Save a measurement, review its cost and repeat it daily, weekly or monthly. Your agent can manage trackers and read their stored results in ordinary conversation.

Tracker lifecycle

1. 1
   Create a free draft

   Save the measurement, prompts or keywords, schedule and selected models. No provider calls start.
2. 2
   Review and activate

   Your agent reads the current cost estimate and asks you to approve that exact maximum per run. Activation enables recurring billed runs.
3. 3
   Monitor or pause

   Read the stored reports for free. Pause to stop future scheduled runs. Editing the definition returns it to Draft for review.

What to provide

Choose one measurement: AI Visibility, AI Competitors, AI Citations or AI Sentiment. Provide one to 30 prompts, the relevant brands or page, selected models, a daily, weekly or monthly schedule and a timezone.

For Keyword Rankings, provide a site and up to 100 keywords instead. Include the Google country, language and device you want to measure.

Need a prompt list first? Start with AI Prompt Research. Trackers can also opt into the ChatGPT surface and measure AI answers from a supported country.

Create a tracker in chat

ChatGPT

Cite42 connected

YouCreate a weekly visibility tracker for Acme CRM. Track Acme CRM, Rival CRM and Legacy CRM on these questions: • best startup CRM • Which startup CRM is quickest to set up? • What CRM should a small sales team use? Run every Monday at 09:00 Europe/London. Show me the estimate and leave it as a draft.

MCPCreate a tracker draft

ChatGPTDraft saved. The 3 prompts across three AI surfaces are estimated at $0.39 per run. Nothing is scheduled yet; approve that maximum before activation.

For the 3-prompt example using three AI surfaces, the current estimate is $0.39 per run. The draft is saved without scheduling any calls. Your agent should show the current estimate before asking you to activate.

After reviewing the current estimate

ChatGPT

Cite42 connected

YouActivate the tracker with a maximum of $0.39 per run.

MCPActivate a tracker

ChatGPTThe tracker is active and will run every Monday at 09:00 Europe/London. Future runs stop if the estimate exceeds your approved maximum.

Once active, the tracker runs on its schedule until paused. A changed estimate requires fresh approval; Cite42 pauses before a run that exceeds the approved maximum.

Available actions

• Create a draft

  Save a tracker and get its estimate. No runs start.

• Update a tracker

  Replace the definition and return the tracker to Draft.

• Activate or resume

  Approve the current per-run maximum and start recurring runs. The balance must cover one run.

• Pause a tracker

  Stop future scheduled runs and keep the configuration and history.

• Delete a tracker

  Remove a Draft or Paused tracker and its history after active work finishes.

• Run now

  Queue one billed manual run with a request token that makes retries safe.

• List trackers

  Read status, schedules, configuration and current cost estimates for free.

• Read a report

  Read stored results and costs for free: eight runs by default, up to 20.

Read the results

Reports contain stored measurements, changes and costs. Reading a report does not rerun the measurement. Each tracker follows one measurement, so create separate trackers if you need both rankings and sentiment history.

ChatGPT

Cite42 connected

YouShow the last 10 runs of my weekly visibility tracker. Summarize the biggest changes and state which answers or keyword checks were unavailable.

MCPRead a tracker report

ChatGPTHere are the stored measurements and changes across the last 10 runs, with unavailable checks called out. Reading this report has not started another measurement.

Billing & confirmation

Management and stored reports are free. Scheduled and manual runs charge for successful saved calls. Activation approves recurring charges at the current maximum per run until paused.

Ask your agent to reread the estimate before activation. Low funds, spend limits or a higher estimate can pause future work. For precise fields, retries and HTTP operations, see the Tracker API reference.