Long tail keywords and questions from Google Autocomplete A-Z for Web, YouTube, Shopping and News, about $0.07 per 1,000 keywords on the Free plan (one topic: 48 checks, $0.048, about 700 keywords). Monitor mode tracks new and lost suggestions across runs.
2026-10-31: Long tail keywords from one topic, lower price
Features
New longTail mode: type a topic in Seeds and get every unique Google
Autocomplete suggestion for it, expanded with A-Z and 0-9, question words,
prepositions or comparison words. One topic with A-Z and questions is 48
checks ($0.048 on the Free plan) and returned about 700 keywords for a US English
Web topic.
Each keyword row shows the prefix that returned it, its rank there, how many
prefixes returned it, word count and whether it is a question. A new
Long tail keywords Dataset view shows these columns.
Question, preposition and comparison words for English, German, French,
Spanish, Portuguese and Italian. Other languages use the alphabet.
depth: 2 expands the top suggestions of each topic again with the alphabet;
maxQueriesPerSeed caps the paid checks per topic, so the most a run can cost
is known before it starts.
The five keywords per topic returned by the most prefixes link to Google
Trends Monitor for a comparison of interest.
Improvements
The input form now opens on long tail mode with one example topic. API calls
that leave out mode still run changesOnly exactly as before, and every
monitor input, output field, status and saved monitor key keeps working.
targets is no longer required in the input schema: a run needs seeds or
targets. An input with seeds and no targets runs long tail mode.
A topic with missing or incomplete keywords gets a status row (PARTIAL,
NO_RESULTS, FAILED, SKIPPED or INVALID_INPUT), so an empty result is
never read as "no keywords". Failed prefixes are never charged.
A long tail run that would outlast its timeout stops starting new topics and
reports them as SKIPPED with RUN_DEADLINE instead of timing out.
If Google stops answering (12 checks in a row fail), a long tail run stops
sending requests; the topics left are reported as FAILED or PARTIAL and
are not charged.
Pricing
Lower price in both modes: a verified prefix check costs $0.001 on the Apify
Free plan instead of $0.002, and less on paid plans: $0.0009 Bronze, $0.0008
Silver, $0.00065 Gold and above. Still no start fee,
and failed checks are still free.
No action is needed for existing monitor runs, Tasks or schedules.
2026-09-24: Results are saved before billing
Fixes
A verified prefix check is charged only after its result rows have been saved
to the Dataset. A failed or unconfirmed write does not trigger a charge.
If the platform declines a charge after the result was saved, that result
remains available without charge. Remaining undelivered checks are skipped,
and RUN_SUMMARY.uncharged_delivered_targets reports the free delivered check.
Interrupted runs still stop without replaying rows or charges. Saved results
remain available; a new run may repeat changes if comparison state was not saved.
No input changes are required. Runs pinned to an older build must switch to the
updated build or latest to receive this fix. The price remains $0.002 per check.
2026-09-09: First release (0.1)
Features
Monitor Google Autocomplete prefixes by country, language and search type:
Web, YouTube, Shopping or News. Results include current suggestions and
NEW_SUGGESTION, LOST_SUGGESTION, RANK_CHANGED and METADATA_CHANGED
events.
Confirm new and missing suggestions across successful checks. A failed
request keeps the previous observation and cannot report a lost suggestion.
Each input has an outcome, including unchanged checks and duplicates. Use
the default view for suggestions and statuses, and the changes view for events.
Five saved examples cover keyword research, market comparisons and monitoring.
The webhook example uses clean=true to omit empty change projections.
Improvements
Changing confirmation or relevance settings starts a new comparison baseline.
Below-threshold suggestions still returned by Google do not produce loss events.
The query echo is checked against the requested prefix. Incomplete or mismatched
responses are reported as failures; a failure across the whole source fails the run.
Spending limits preserve comparison state. Interrupted delivery is checked
before resuming to avoid repeating rows or charges.
Google relevance scores appear in the output but do not trigger change events.
Autocomplete positions do not measure search volume or website rankings.
Pay per successful prefix check. Retries, failed checks and duplicate scopes
are not charged. Access errors such as 401, 403 and 451 remain source failures,
so they do not imply that a keyword has no suggestions.
