Tracker tools

Create weekly or monthly trackers, confirm their per-run cost before activation, run them manually, read stored results, and permanently remove inactive trackers through eight MCP tools.

Tracker lifecycle

A tracker is one measurement, rankings, compare, citations, or sentiment, applied to a list of prompts on a weekly or monthly cadence. The brand, competitors and AI surfaces are set once for every prompt, and each prompt is one billed call per run. Your agent can create and review the draft for free before any recurring calls begin.

The draft returns estimatedCostMicroPerRun. Activation requires that exact value as maxCostMicroPerRun, so the current maximum is explicitly accepted before the schedule starts. If the estimate changes, activation is rejected until the new amount is reviewed and confirmed. Recurring calls continue until the tracker is paused.

No prompt list yet? cite42_map_ai_prompts returns exactly what cite42_tracker_create takes, so an agent can research the prompts worth watching and then put the best of them on a schedule in the same conversation.

Create a tracker in chat

Ask for the tracker in plain English and tell your agent to keep it as a draft. This first turn saves the configuration and returns the exact current estimate without starting any scheduled or manual runs.

# you say
Create a weekly visibility tracker for Acme CRM.
Name it “Acme CRM weekly visibility”.
Track rankings for Acme CRM, Rival CRM, and Legacy CRM on these prompts:
- 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 cost, but do not activate it.

cite42_tracker_create returns the saved draft to your agent:

{
  "tracker": {
    "id": "6d9d6580-4cef-4b84-8804-2e90d9e57c9a",
    "name": "Acme CRM weekly visibility",
    "status": "DRAFT",
    "cadence": {
      "frequency": "weekly",
      "timezone": "Europe/London",
      "time": "09:00",
      "weekday": 1
    },
    "primitive": "rankings",
    "prompts": [
      "best startup CRM",
      "Which startup CRM is quickest to set up?",
      "What CRM should a small sales team use?"
    ],
    "brand": null,
    "competitors": [
      "Acme CRM",
      "Rival CRM",
      "Legacy CRM"
    ],
    "url": null,
    "models": [
      "chatgpt",
      "perplexity",
      "gemini"
    ],
    "estimatedCostMicroPerRun": 39000,
    "maxCostMicroPerRun": null,
    "nextRunAt": null,
    "lastRunAt": null,
    "pauseReason": null,
    "createdAt": "2026-08-17T08:30:00.000Z",
    "updatedAt": "2026-08-17T08:30:00.000Z"
  }
}

Your AI explains the result and asks for the separate billing confirmation:

# your AI replies
Draft created. 3 prompts across three AI surfaces, so the current estimate is 39,000 micro-credits ($0.39) per run.
Nothing is scheduled yet. Approve that exact amount as the maximum per run to activate it.

Activation is a second, explicit turn:

# you say
Activate Weekly AI visibility with a maximum of 39,000 micro-credits per run.

# MCP calls cite42_tracker_activate
{
  "trackerId": "6d9d6580-4cef-4b84-8804-2e90d9e57c9a",
  "maxCostMicroPerRun": 39000
}

# your AI replies
The tracker is Active. It will run every Monday at 09:00 Europe/London while active.
If a future estimate exceeds the approved maximum, Cite42 pauses it before another provider call.

The eight tools

  • cite42_tracker_create: create a free draft from a name, cadence, one measurement and a list of prompts. It does not schedule calls.
  • cite42_tracker_update: replace the whole definition for free, a tracker read returns the same shape, so read it, change what you need, and send it back. Any edit returns it to draft so the updated recurring maximum must be confirmed before scheduling resumes.
  • cite42_tracker_activate: activate or resume recurring runs after confirming the latest per-run maximum. The response includes the balance, the cost per run, and runsDryAt, the date the balance stops covering the account’s active trackers. It is refused with 402 when the balance cannot cover even the next run.
  • cite42_tracker_pause: stop future scheduled calls for free while keeping configuration and run history.
  • cite42_tracker_delete: permanently remove a draft or paused tracker and its run history. Active trackers must be paused first, and deletion is rejected while a run is queued or running. Financial usage records are retained.
  • cite42_tracker_run: queue one manual run from the saved inputs. Use a new request token for each intended run and reuse it only to retry that same run.
  • cite42_trackers: list tracker status, cadence, configuration, estimate, accepted maximum, and next/last run times for free.
  • cite42_tracker_report: read bounded stored run history, including per-primitive results, changes, statuses, sampled times, and costs, for free. It returns the latest 8 runs by default and accepts up to 20 per call.

Billing & confirmation

Creating, updating, activating, pausing, deleting, listing, and reading stored tracker history cost 0 credits. Each scheduled or manual run bills the saved primitive calls at their normal per-model rates. Activation itself is free, but it authorizes those future billed calls until pause; that is why cite42_tracker_activate requires the current estimate to be confirmed exactly.

Use cite42_pricing for the current rate table and cite42_trackers to re-read a tracker’s estimate before activating it.

More prompts

Ask in plain language; your agent coordinates the lifecycle tools:

# create and review a draft without starting calls
Create a weekly Cite42 tracker for our rankings, competitor comparison, citations, and sentiment. Show me its credits-per-run estimate, but do not activate it.

# explicit confirmation before recurring billing
Re-read that tracker's current estimate and ask me to approve the exact maximum before you activate its recurring runs.

# free stored-history read
Show me the last 10 runs for our weekly visibility tracker and summarize the biggest ranking, citation, and sentiment changes.