Skip to main content
CreatorAudit collects data on a recurring schedule, not in real time. When you start tracking an account or a video, CreatorAudit fetches it and then refreshes it periodically — so values reflect the most recent collection, not the live platform.
Treat the API as a stream of eventually-consistent snapshots. A value won’t appear the instant you create a resource, and it may lag the platform by a refresh cycle. Build on the timestamps the API gives you rather than assuming “now”.

​
The first refresh

A newly tracked account or video has no metrics until its first collection completes.
  • In the API, the account or video’s last_scrape_time is null until its first refresh; added_at records when you started tracking it.
  • In the dashboard, a new account shows a Scraping… state, and metrics appear once the first refresh lands.
To tell when data is ready, poll the resource and wait for last_scrape_time to become non-null:
1

Create the resource

POST /v2/accounts (or POST /v2/videos). The response includes the new id and added_at, with last_scrape_time still null.
2

Poll until the first refresh lands

Re-fetch GET /v2/accounts/{id} on a sensible interval (seconds to minutes, not a tight loop) until last_scrape_time is set. Then its metrics and videos are populated.

​
Refresh cadence

Tracked accounts — and the videos discovered in their feeds — are refreshed on a recurring cycle. For an individually tracked video, you can set scrape_interval_hours to control how often CreatorAudit re-fetches that specific video.
Don’t hardcode a fixed refresh frequency. The exact cadence depends on your plan and your per-account and organization scraping settings (see Usage & billing). Read last_scrape_time to know how current a value is.

​
Time series fill in over time

Daily analytics accrue across successive refreshes. A brand-new account or video has little history, and its time series fills in over the following days as more snapshots are collected. Expect a short window after you start tracking before a 7-, 30-, or 90-day chart is meaningful.

​
Guidance for integrations and agents

  • Poll, don’t busy-wait. After a write, check back on an interval; don’t expect data immediately. There are no webhooks — polling is the supported pattern; see Realtime & webhooks.
  • Anchor on timestamps. Use last_scrape_time and the timestamps in time-series responses to reason about staleness, and cache accordingly.
  • Be patient with new resources. Give a newly tracked account or video time to complete its first refresh before reporting on it.

Read analytics

What each metric means and how windows and time series work.