Scraper Output Guard

Catch silent scraper failures. A run can finish with status SUCCEEDED and still be broken: it returned 40 items instead of 500, a key field went empty, or the site changed and a field disappeared. Scraper Output Guard checks the content of your own Apify Actor runs, not just their status, and gives each Actor a verdict: ok, warning or unhealthy, with the reasons.

What is Scraper Output Guard? An Apify Actor monitor for empty, truncated and drifting data

For each Actor you list, it compares its most recent runs and flags:

• Failed runs: the latest run failed, or the recent failure rate is high
• Item-count drops: fewer items than your minimum, or sharply below the median of recent runs
• Early stops: the run asked for N items (maxItems, maxResults, resultsLimit...) but returned far fewer while reporting success
• Empty required fields: fields you name (e.g. price or seller.name) are empty in too many sampled items
• Schema drift: fields present in an older run's first sampled item are missing from the latest run's first sampled item
• Duplicates: repeated items in the sample

Who it is for

Anyone who runs scrapers on a schedule and feeds the data into reports, price monitors, lead lists or AI pipelines, where a quietly half-empty dataset costs more than a failed run.

Try it free first

Run it with no input. It runs a free demo on built-in sample data: three sample Actors (one healthy, one truncated, one with schema drift) so you can see the verdicts and output format. Demo runs are not billed. One demo result:

{
  "actorId": "demo/truncated-scraper",
  "verdict": "unhealthy",
  "reasons": [
    "Item count 40 is more than 50% below the recent median 500.",
    "Input maxItems=500 but only 40 items (possible early stop)."
  ],
  "latestRunId": "d4",
  "latestItemCount": 40,
  "runsChecked": 3,
  "demo": true
}

How to use it on your Actors

1. In Apify Console, create an API token with read access to the Actors, runs and datasets you want checked.
2. Paste it into Apify API token, and list Actor IDs (username/actor-name, up to 50).
3. Optionally list required fields as dotted paths, and adjust the thresholds. Example input: {"apiToken": "<read-only token>", "actorIds": ["username/my-scraper"], "requiredFields": ["price"]}
4. Schedule it (for example daily after your scrapers run). Open the dataset's Verdicts view. The SUMMARY record in the key-value store has the totals. Turn on Only output Actors with problems to keep the dataset short.

Run it automatically: schedules, webhooks, Make, Zapier, n8n

• Use an Apify schedule to run it daily after your scrapers.
• Or start it from a webhook when a scraper run finishes.
• Like any Actor, it can be started from Make, Zapier or n8n through Apify's integrations, so its verdicts can feed your alerts.

Apify Actor run monitor: when to use this

• Status alerts only: if you only need alerts when runs fail, Apify's free monitoring Actors may be enough.
• Content checks: use Scraper Output Guard when a run can "succeed" with bad data. It combines item-count drops, early stops, empty required fields, schema drift and duplicates into one verdict per Actor, from a single run with no extra setup.

Output

One dataset item per Actor checked: actorId, verdict (ok, warning, unhealthy, or unknown if it couldn't be read), reasons, latestItemCount, latestRunId, runsChecked. Reasons may name fields and counts; your item values are never output.

How much does it cost to monitor Apify Actors?

Pay per event: $0.005 per Actor checked (actor-checked event).

• Example: checking 10 Actors once a day is about 300 events a month, which is $1.50.
• Billed: every Actor that gets a verdict (ok, warning or unhealthy, including "No runs found").
• Not billed: Actors that can't be read (unknown); demo runs.
• Cap: set a maximum charge per run in the run options.
• See the Pricing tab for the full, current price.

Privacy and limits

• Your token is sent only to api.apify.com. It is never logged or written to the output. Use a read-only token.
• Reads run metadata and up to 1,000 sampled items per run.
• These are heuristic checks. They can't prove your data is correct or complete.
• This is an unofficial utility, not affiliated with Apify. It was built with the help of AI tools under the developer's direction, and its checks were tested on sample data.

FAQ

Does it re-run or change my Actors? No. It only reads runs and datasets.

What if a run is still in progress? The Actor gets a warning that says so. Schedule the guard after your scrapers usually finish.

How do I monitor Apify Actors for empty datasets? List them in Actor IDs, set Minimum items per successful run, and schedule this Actor daily. Any run below the minimum is flagged unhealthy.

Can I check Actors built by other developers? Yes, if your token can read your runs of them. Apify's own test only checks that an Actor returns at least one item; this checks the size and content of your runs.

Why does it need a token? To read your runs and datasets through the Apify API. Use a read-only token; it is never logged or written to the output.