A pricing change notice arrived on July 28, 2026 for one YouTube transcript actor. Reading it properly took ten minutes, and the arithmetic was interesting enough that we audited the whole category the same afternoon. Eight actors doing one job. The prices differ by 500 times, the reliability differs by 15 percentage points, and the two are not related.
Apify's pay-per-event model charges separately for starting an actor and for each result it produces. This change moved both numbers in opposite directions, which is why a single percentage cannot describe it. The start fee fell by 95%, and the per-transcript charge rose. So the answer to "cheaper or dearer" depends entirely on how many transcripts you pull per run.
| What you pay for | Before | After | Change |
|---|---|---|---|
| Actor start, per run | $0.001 | $0.00005 | −95% |
| Per transcript, paid plans | $0.0005 | $0.0007 | +40% |
| Per transcript, free plan | $0.0005 | $0.001 | +100% |
| A 1-transcript run, paid plan | $0.0015 | $0.00075 | −50% |
| A 1,000-transcript run, paid plan | $0.501 | $0.70005 | +40% |
| A 1,000-transcript run, free plan | $0.501 | $1.00005 | +100% |
supreme_coder/youtube-transcript-scraper · read from the live pricing record on July 28, 2026
One more detail worth knowing: the actor's own title advertises $0.7 per 1k, and that rate is real on paid Apify plans. On the free plan the same actor charges $0.001 per transcript, which is $1.00 per 1,000. Tiered pay-per-event pricing is normal on Apify, but the headline number is the one your plan does not necessarily get.
One repriced actor tells you nothing about whether it is a good choice. So we pulled the eight most-used YouTube transcript actors on the Apify Store and audited all of them with Audit Tools, in free metadata mode, on the afternoon of July 28, 2026.
Audit Tools scores actors on 23 criteria. Free metadata mode measures the subset that live platform data can prove without spending money, which was eight criteria here: 30-day marketplace reliability, pricing transparency, rot check on last modified date, listing freshness, publisher pulse, permission posture, agent payability and adoption trajectory. Every score carries an evidence sentence, so each number below can be traced back to a count of real runs. Criteria that require an actual run, such as cost per usable item and end-to-end latency, are measured in a funded deep scan instead. This audit cost nothing to run.
| Actor | Score | Per 1,000 | 30-day success | Runs (30d) | Rating | Updated |
|---|---|---|---|---|---|---|
| YouTube Transcript Scraper Pro codepoetry/youtube-transcript-ai-scraper |
97/100 | $1.00 | 97.8% | 47,730 | no reviews | 21 days |
| YouTube Video Transcript most reliable starvibe/youtube-video-transcript |
96/100 | $5.00 | 100.0% | 492,138 | 5.00 (20) | 122 days |
| Youtube Transcript Scraper supreme_coder/youtube-transcript-scraper |
96/100 | $1.00 | 99.0% | 106,364 | 5.00 (2) | 14 days |
| YouTube Transcript & Subtitles Scraper API johnvc/YoutubeTranscripts |
96/100 | $0.02 | 95.5% | 9,394 | 5.00 (5) | today |
| Youtube Transcripts karamelo/youtube-transcripts |
95/100 | $7.00 | 99.9% | 60,664 | 4.99 (15) | 32 days |
| YouTube Transcript Ninja topaz_sharingan/Youtube-Transcript-Scraper-1 |
95/100 | $10.00 | 99.7% | 38,937 | 5.00 (28) | 9 days |
| Youtube Transcript Scraper pintostudio/youtube-transcript-scraper |
88/100 | $10.00 | 99.7% | 378,037 | 4.78 (46) | 145 days |
| Best Youtube Transcripts Scraper scrape-creators/best-youtube-transcripts-scraper |
88/100 | $1.00 | 85.0% | 46,111 | 4.14 (12) | 97 days |
Free-plan cost for 1,000 transcripts · 30-day success = succeeded ÷ total runs · Audit Tools metadata mode, July 28, 2026. Per-1,000 figures assume one transcript per result. Tiered actors are cheaper on paid plans: karamelo drops to $5.00 and supreme_coder to $0.70. The Pro actor bills AI fallback transcription separately at $0.012 per minute on the free plan, charged only when captions are missing, so its real cost depends on how many of your videos lack captions.
High volume, production pipeline, reliability first: YouTube Video Transcript (starvibe/youtube-video-transcript). 492,138 runs in 30 days with 5 failures. Nothing else in the category has that combination of volume and cleanliness. At $5.00 per 1,000 it is mid-priced, and 122 days since the last update is the one thing to watch.
Cost-sensitive bulk work where a retry is cheap: YouTube Transcript and Subtitles Scraper API (johnvc/YoutubeTranscripts). About $0.02 per 1,000 transcripts, roughly 250 times cheaper than the dearest option, scored 96, and updated the day we audited it. The trade is maturity: 9,394 runs, 5 reviews, and 322 timeouts to plan retries around. A price change is also dated July 30, 2026.
Videos that have no captions at all: YouTube Transcript Scraper Pro (codepoetry/youtube-transcript-ai-scraper). Highest score of the eight at 97, and the only one that falls back to AI transcription when captions are missing. That fallback is metered by the minute on top of the per-transcript fee, so model your caption-miss rate before committing. No reviews yet.
Channels, playlists and search results in one run: Youtube Transcript Scraper (supreme_coder/youtube-transcript-scraper). The July 28 update added exactly this, and reliability is solid at 99.0% across 106,364 runs. Just price it at your real batch size: on the free plan any run of two or more transcripts now costs more than it did before the change.
Long-standing option with the deepest review history: Youtube Transcript Scraper (pintostudio/youtube-transcript-scraper). 46 reviews at 4.78 and 378,037 runs at 99.7% success, so it works. It scored 88 because it is the stalest in the group at 145 days since the last update, and at $10.00 per 1,000 it is also the joint most expensive.
Two of these eight actors change price inside one week. supreme_coder repriced on July 28, 2026. johnvc has a change dated July 30, 2026, two days after we published this. Looking further back, pintostudio has repriced five times since November 2024 and johnvc eight times since November 2025. Every table above is true on July 28, 2026 and carries an expiry date we cannot see. That is the whole argument for auditing before you run rather than trusting a comparison someone wrote months ago, including this one.
Q. What is the best Apify YouTube transcript scraper in 2026?
We audited eight of them on July 28, 2026. For reliability at volume, YouTube Video Transcript (starvibe/youtube-video-transcript) is the strongest pick: 96/100, with 492,119 of 492,138 runs succeeding over 30 days, a 100.0% success rate with only 5 failed runs, at $5.00 per 1,000 transcripts on the free plan. The highest raw score was YouTube Transcript Scraper Pro (codepoetry/youtube-transcript-ai-scraper) at 97/100, but it has no reviews yet and meters AI fallback transcription by the minute on top of the per-transcript price.
Q. How much does it cost to scrape 1,000 YouTube transcripts on Apify?
Anywhere from about $0.02 to $10.00 for the identical job, depending which actor you pick. Free-plan cost for 1,000 transcripts, as measured on July 28, 2026: johnvc/YoutubeTranscripts about $0.02; codepoetry, supreme_coder and scrape-creators $1.00; starvibe/youtube-video-transcript $5.00; karamelo/youtube-transcripts $7.00; pintostudio and topaz_sharingan $10.00. That is a 500x spread for the same output, and in this category the price told us nothing useful about the reliability.
Q. Why did Youtube Transcript Scraper pricing change to $0.7 per 1k in July 2026?
The author repriced supreme_coder/youtube-transcript-scraper on July 28, 2026, giving the reason that it now supports scraping full channels, search results and playlists. The actor start event fell from $0.001 to $0.00005, and the per-transcript charge rose from $0.0005 to $0.0007 on paid plans and $0.001 on the free plan. The effect depends on batch size: a single-transcript run got about 50% cheaper, while a 1,000-transcript run went from $0.501 to $0.70005 on a paid plan, about 40% more. The crossover is five transcripts per run on paid plans and two on the free plan.
Q. Is the $0.7 per 1k YouTube transcript price the same on the Apify free plan?
No. The $0.70 per 1,000 rate applies on paid Apify plans. On the free plan that actor charges $0.001 per transcript, which is $1.00 per 1,000, so free-plan users pay double the headline rate in the actor's own title. Tiered pay-per-event pricing is standard on Apify and several actors in this category use it. It is worth reading your own plan's rate rather than the number in the listing name.
Q. Which Apify YouTube transcript actor is the most reliable?
YouTube Video Transcript (starvibe/youtube-video-transcript) had the cleanest 30-day record of the eight: 492,119 succeeded out of 492,138 runs, with 5 failed, 7 timed out and 7 aborted, a 100.0% success rate on the largest volume in the group. Youtube Transcripts (karamelo/youtube-transcripts) was next at 99.9%, then pintostudio and topaz_sharingan at 99.7%. The weakest was Best Youtube Transcripts Scraper (scrape-creators/best-youtube-transcripts-scraper) at 85.0%, with 6,659 failed runs out of 46,111.
Q. Is the cheapest YouTube transcript actor on Apify any good?
The cheapest one we audited scored well but has the thinnest track record. YouTube Transcript and Subtitles Scraper API (johnvc/YoutubeTranscripts) works out to roughly $0.02 per 1,000 transcripts and scored 96/100, with the freshest code in the group, updated the same day we ran the audit. The catch is volume and timeouts: 9,394 runs in 30 days with zero failures but 322 timeouts, a 95.5% success rate, and only 5 reviews. Cheap and current, but far less proven than actors with hundreds of thousands of runs behind them. It also has a price change dated July 30, 2026.
Q. Does an actor named 'best' actually perform best on Apify?
Not in this category. Best Youtube Transcripts Scraper (scrape-creators/best-youtube-transcripts-scraper) finished last of the eight actors we audited on July 28, 2026. It had the lowest 30-day success rate at 85.0%, meaning roughly one run in seven failed, the lowest community rating at 4.14 out of 5, and it tied for the lowest overall score at 88/100. An actor name is listing copy written by its author. It is not a measurement, and it takes about a minute to check.
Q. Why does an actor with zero failed runs still fail my pipeline?
Because timeouts and aborts are counted separately from failures. Two actors in this audit reported zero failed runs in 30 days yet still lost runs: johnvc/YoutubeTranscripts had 0 failed but 322 timed out of 9,394 runs, and topaz_sharingan/Youtube-Transcript-Scraper-1 had 0 failed but 71 timed out and 32 aborted. A vendor can truthfully claim no failures while your job returns nothing. Read succeeded divided by total runs instead of the failure count on its own.
Q. How often does Apify actor pricing change?
Often enough that a hardcoded choice goes stale. Two of the eight YouTube transcript actors we audited changed price inside one week: supreme_coder/youtube-transcript-scraper on July 28, 2026 and johnvc/YoutubeTranscripts with a change dated July 30, 2026. Across their full histories, pintostudio/youtube-transcript-scraper has repriced five times since November 2024 and johnvc/YoutubeTranscripts eight times since November 2025. Several actors also migrated from per-dataset-item pricing to pay-per-event in early 2026. Any actor you picked six months ago may cost a different amount today for the same call.
Q. How does Audit Tools score an Apify actor, and what does it cost?
Audit Tools (audit-tools.ai) scores actors on 23 criteria, including 30-day marketplace reliability, pricing transparency, rot check on last modified date, listing freshness, publisher pulse, permission posture, agent payability and adoption trajectory. Free metadata mode measures what live platform data can prove without spending money, which was eight criteria in this audit, and every score carries an evidence sentence naming the counts behind it. Criteria that need a real run, such as cost per usable item and end-to-end latency, are measured in a funded deep scan instead. The eight-actor audit in this post ran in free metadata mode and cost nothing.
Prices move, authors go quiet, and success rates drift. The table above was true on the afternoon of July 28, 2026, and one of the actors in it already has a price change queued for two days later. Audit Tools scores an actor as it exists at the moment you ask, which is the only version that can break your pipeline. Metadata mode is free.
Measured, not marketed.