Run Tracking
Understand queued, running, measured, and failed tracking states.
Tracking runs collect AI answers for active prompts across active engines and markets.
Run tracking now
The primary user-facing action is Run tracking now. It queues work and returns quickly while background workers process AI answers.
What gets queued
For the primary workflow, planned runs are:
active prompts x active plan-allowed engines x active markets x repeat countThe active engine set is plan-gated. Starter, Growth, and current agency self-serve plans include ChatGPT, Gemini, Perplexity, DeepSeek, and ByteDance. Brand Pro and Custom plans can include Grok.
The default repeat count is five. Repeat counts may be lower or higher only when an internal measurement workflow explicitly configures them within the allowed range.
Weekly Google SERP tracking
Separately from the daily runs, every plan checks Google AI Overviews and
Google AI Mode once per ISO week for each active prompt and location, with a
single run per check (Google's SERP output is far less variable than model
answers, so repeat sampling adds little). These runs record whether the
surface appeared for the query (surface_present), the answer text, and the
sources it cites. They attach to that day's measurement window and never count
against the daily prompt allowance.
Statuses
| Status | Meaning |
|---|---|
| Not measured yet | The prompt has been saved, but no answer is available. |
| Queued | The run has been planned and is waiting for processing. |
| Running | The provider request is in progress. |
| Measured | A completed answer is available. |
| Needs attention | The run failed or requires review. |
Resume behavior
If queued or running jobs already exist, Citlyze reuses the active measurement window instead of creating duplicate work.
Background processing
Processing happens outside the page request. This keeps large workspaces from timing out while AI providers return answers.