Audio & Video Transcriber: Podcasts, SRT, Speakers

What does Audio & Video Transcriber do?

Audio & Video Transcriber turns audio and video into text: upload a file, paste media links, or point it at a podcast RSS feed. You get timestamped segments, speaker labels, the detected language and ready-made SRT and WebVTT subtitles, plain text and JSON. It also takes TikTok, X (Twitter), Facebook, SoundCloud, Dailymotion and YouTube links, Apple Podcasts links and Google Drive, Dropbox and OneDrive share links.

The speech recognition runs at Deepgram, AssemblyAI or Groq, not inside the Actor, so it is fast and accurate. You are charged per second of audio, and failed or silent media are free.

Who is it for?

• Podcasters and podcast researchers: transcripts and subtitles for new episodes, on a schedule, without paying twice for the same episode.
• Video creators and agencies: SRT and VTT subtitle files for videos, with speaker labels for interviews.
• Journalists, analysts and lawyers: searchable text from recordings, calls and public videos.
• Developers: a clean API for "media URL in, text out", with a dataset hand-over from other Apify scrapers (TikTok, X, Facebook ...).

Two ways to pay: managed or bring your own key

	Managed (default)	BYOK (bring your own key)
Who can use it	Paid Apify plans	Every plan, including free
Provider keys	The Actor owner's	Your own (Deepgram, AssemblyAI, or OpenAI-compatible such as Groq or OpenAI)
Price	$0.000167 per second ($0.01 per minute)	$0.0000667 per second ($0.004 per minute) plus your provider's own bill
Engines	Deepgram nova-3, then AssemblyAI, then Groq whisper-large-v3-turbo; the next one takes over when one runs out of credit	The one you choose

Managed uses the owner's provider accounts and is available on paid Apify plans. Set the prices in the Apify Console; they are not hard-coded.

On a free plan, managed mode returns one free info row that tells you to switch to BYOK (the run still succeeds). Choose Mode: BYOK, a provider and your key, and it works.

What you can give it

• Upload one file (mediaFile), or up to 20 media URLs (mediaUrls): direct audio or video files, platform links and share links.
• Podcast RSS feeds (podcastFeeds): the newest maxEpisodesPerFeed episodes (default 3, up to 50), with publishedAfter and episodeTitleContains filters. Each row carries the episode title, publication date and podcast name.
• Apple Podcasts links: resolved to the podcast's RSS feed through Apple's lookup API. Spotify links: Spotify audio is DRM protected and is never downloaded; the Actor looks for the same podcast's public RSS feed and, if it cannot find it, returns UNSUPPORTED_SOURCE (free).
• Google Drive, Dropbox, OneDrive and SharePoint share links: turned into direct download links.
• Another Actor's dataset (datasetId or pasted datasetItems): the media URL field (videoUrl, mediaUrl, audioUrl, downloadUrl, url ...) is found automatically. Good after a TikTok, X or Facebook scraper.
• Social links (TikTok, X/Twitter, Facebook, SoundCloud, Dailymotion, YouTube, Reddit, Instagram, Vimeo): only the audio is downloaded with yt-dlp, up to maxMinutesPerItem (default 120).

Social platforms: what works from a data center

Tested on 7 October 2026 from a data center without a proxy: TikTok, X/Twitter, Facebook, SoundCloud and Dailymotion work. YouTube and Reddit block data centers; Instagram and Vimeo want a login. For those:

• Cookies (cookies, Netscape format, stored encrypted, never logged or saved): paste your browser cookies and logged-in or blocked links usually work.
• Residential proxy fallback (paid plans only): if a platform blocks the download, the Actor makes one more attempt through an Apify residential proxy and charges platform-proxy-second ($0.00025 per second, $0.015 per minute) on top, only if it worked. Free plans never use the proxy; a blocked link returns PLATFORM_BLOCKED with a hint to use cookies. The owner can switch the fallback off in the Console (PROXY_FALLBACK_ENABLED other than 1) without a new build.

Output

One dataset row per media, plus files in the key-value store.

Row field	Meaning
status	completed, empty (silence, free), failed (free) or info (a free note)
text, segments	The transcript and {start, end, text, speaker} segments in seconds
language	The language you set, or the one detected
durationSeconds, chargedSeconds, proxySeconds	Length and what was charged
title, publishedAt, podcastName	For podcast episodes and platform links
words	Per-word timestamps, with wordTimestamps on
speakerLabelsUnavailable	True if you asked for speaker labels but only Groq (which cannot label speakers) was available
srtKey, vttKey, txtKey, jsonKey	Keys of the files in the key-value store (transcripts_<n>.srt ...)
error, errorMessage	A short code and a plain explanation

In SRT and VTT, speakers appear as Speaker 1: ..., Speaker 2: ....

Error codes: invalid-url, blocked-url, download-failed, file-too-large, ffmpeg-failed, duration-failed, unsupported-option, provider-http-error, provider-timeout, provider-response-invalid, billing-failed, output-failed, unexpected-error, MANAGED_UNAVAILABLE, PROVIDER_CREDIT_EXHAUSTED, BUDGET_LIMIT, UNSUPPORTED_SOURCE, PLATFORM_BLOCKED, LOGIN_REQUIRED, MAX_DURATION_EXCEEDED.

How much does it cost?

This Actor uses pay per event, counted in seconds (the length of the media, rounded up):

Event	Price	When
transcript-second-managed	$0.000167 per second ($0.01 per minute)	Managed mode, successful non-empty transcript
transcript-second-byok	$0.0000667 per second ($0.004 per minute)	BYOK mode, successful non-empty transcript
platform-proxy-second	$0.00025 per second ($0.015 per minute)	Only when a platform link needed the residential proxy and was transcribed

• A 30-minute podcast episode costs $0.30 managed or $0.12 BYOK (plus your provider's bill in BYOK).
• Failed, empty and info rows are free, and so is anything that cannot be finished: media over the length limit, a blocked platform, a provider out of credit, a link that is not media.
• Budget check (paid plans): before each media the Actor checks that it fits in your Maximum cost per run. If not, it skips it with BUDGET_LIMIT and no charge.
• Never charged twice: use a watchlist (watchlistId + newItemsOnly) so scheduled runs only transcribe media they have not done before.

Other options

Field	Meaning
language	ISO-639-1 code (en, tr, de). Empty = detect automatically; the detected language is in every row.
speakerLabels	Deepgram and AssemblyAI only; no extra charge.
translateToEnglish	BYOK with an OpenAI-compatible provider only (Whisper translation).
wordTimestamps	Word-level times in the row and JSON file; no extra charge.
outputFormats	text, srt, vtt, json (default all four).
maxMinutesPerItem	Longer media is skipped for free (default 120).

Input example

BYOK with a Groq key (a free Groq key works on a free Apify plan too):

{
    "mode": "byok",
    "byokProvider": "openai_compatible",
    "byokBaseUrl": "https://api.groq.com/openai/v1",
    "byokModel": "whisper-large-v3-turbo",
    "byokApiKey": "<your key, entered in the secret field>",
    "podcastFeeds": ["https://feeds.simplecast.com/Sl5CSM3S"],
    "maxEpisodesPerFeed": 2,
    "watchlistId": "the-daily",
    "newItemsOnly": true
}

Managed (paid plans):

{ "mode": "managed", "mediaUrls": ["https://soundcloud.com/forss/flickermood"], "speakerLabels": true }

Never put a key in a file that you commit; enter it in the secret field.

What a local test returned

On 7 October 2026 the Actor was run locally (BYOK, OpenAI-compatible, with a local stand-in for the provider, because no provider key was available), with these links:

Link	Result
SoundCloud track (Forss, "Flickermood")	Completed: audio downloaded with yt-dlp (title, 214 s), converted with ffmpeg, sent to the provider; 214 seconds charged.
Podcast RSS feed, newest episode (The Daily, 22 min 35 s)	Completed: episode picked from the feed, MP3 downloaded and converted; 1,356 seconds charged; title, date and podcast name in the row.
A TikTok video	PLATFORM_BLOCKED: TikTok answered "Your IP address is blocked from accessing this post" from the test machine; no charge. (On Apify, paid plans would try the residential proxy once.)
http://localhost/a.mp3	blocked-url (private addresses are refused).

The transcript text in that run came from the stand-in, so it says nothing about recognition quality. SRT, VTT, text and JSON files were written as expected.

Limitations

• Recognition quality is the provider's. Accents, noise and overlapping speech affect it; speaker labels are best-effort.
• Speaker labels need Deepgram or AssemblyAI. Groq (the last managed fallback) cannot label speakers.
• Long files: media up to 2 GB and maxMinutesPerItem (default 120 minutes). Providers with a 25 MB limit (Groq, OpenAI) get the audio cut into parts, sent one after the other and joined with the right timestamps.
• Spotify audio is never downloaded (DRM); only the same podcast's public RSS feed.
• Platform links change often; yt-dlp is updated with each build. YouTube and Reddit usually need cookies or the proxy fallback; Instagram and Vimeo need cookies.
• Private or login-only share links (Drive, Dropbox, OneDrive) cannot be downloaded; share them with "anyone with the link". Google Drive files above its virus-scan size show a web page instead of the file and return UNSUPPORTED_SOURCE.
• Not legal advice on copyright: only transcribe media you have the right to use.

FAQ

Do I need an API key?

In managed mode (paid plans) no. In BYOK mode yes, from Deepgram, AssemblyAI, Groq or OpenAI. Free plans use BYOK.

Why did I get a row with status "info" and nothing was transcribed?

Managed mode is only for paid plans (or it is not set up). Switch to BYOK and add your key. The info row and the run are free.

Am I charged for silence or failures?

No. Only successful, non-empty transcripts are charged.

Can it transcribe a YouTube video?

Sometimes. YouTube blocks data centers, so the Actor needs your cookies or, on paid plans, the one residential proxy attempt (extra platform-proxy-second charge if it works).

Does it keep my keys?

No. Keys and cookies are never logged and never written to rows or files.

For the Actor owner

Set DEEPGRAM_API_KEY, ASSEMBLYAI_API_KEY and GROQ_API_KEY as secret environment variables in the Apify Console (any one is enough), and enter the three event prices under Publication → Monetization (see docs/PUBLISH.md). APIFY_USER_IS_PAYING is set by Apify. PROXY_FALLBACK_ENABLED is 1 in actor.json; change it in the Console to turn the proxy fallback off without a new build.

Local development:

npm install
npm test
npm run build
ACTOR_TEST_PAY_PER_EVENT=true APIFY_USER_IS_PAYING=1 npm run start:dev

More tools from the same developer

All pay-per-result, no proxy or login needed, built and maintained by the same developer:

Website audits

• Website Accessibility Checker: WCAG 2.2 & EAA: accessibility issues with fixes, SEO basics and security headers.
• Cookie & Tracker Audit: GDPR Consent Checker: cookies and tracking tags that load before consent.
• AI Crawler Access Checker: robots.txt & llms.txt: which AI crawlers a site allows, plus llms.txt.
• Structured Data Validator: Schema & Rich Results: JSON-LD and Microdata checked against Google's rich result rules.
• Website Change Monitor: Diffs, Prices & Alerts: get a row only when a page changes, with a clean diff.
• Policy Change Monitor: Privacy & Terms Tracker: privacy policy and terms changes, classified clause by clause.
• SaaS Pricing Monitor: Track Competitor Plans: competitor pricing plans as structured data, with price-change diffs.
• Sitemap Change Monitor: New & Removed URLs: new, removed and changed URLs from a site's sitemaps.
• Site SEO Crawler: Broken Links & On-Page Audit: site-wide crawl for broken links, redirects and on-page SEO issues.
• Website Tech Stack Detector: Bulk Lead Filters: technologies used by websites, with filters for lead lists.
• Domain Availability & Expiry Checker (RDAP): domain availability, registrar and expiry dates from RDAP.
• Status Page SLA Monitor: Uptime & Incidents: SaaS vendor uptime, incidents and a monthly SLA table from public status pages.

Company data and compliance

• Company Contact Finder: Emails, Phones & Socials: contact details published on company websites.
• UK New Companies Feed: Companies House Daily: newly incorporated UK companies with sector filters.
• EU VAT Number Validator: Bulk VIES Checker: bulk VAT checks with name, address and consultation number.
• LEI Corporate Tree: GLEIF Parents & Subsidiaries: LEI lookup with parents, subsidiaries and a KYC summary.
• UK Company Distress Watch: Insolvency Alerts: administrations, liquidations and winding-up petitions from The Gazette.
• IBAN Validator: Bulk Check & Bank Lookup: IBAN format and checksum, plus bank name and BIC where open data exists.
• Sanctions Screening: OFAC, EU, UN & UK Lists: names screened against official sanctions lists, with match scores.
• OSHA & EPA Violation Monitor: Fines by Company: US OSHA and EPA enforcement cases and penalties by company.
• Email List Precheck: Syntax, MX & Disposable Check: cheap email list hygiene before a real verification.
• UK Company Officer Changes: Director & PSC Alerts: director, owner, address and status changes for a watchlist of UK companies.
• FDA Warning Letters Monitor: By Company & Subject: FDA warning letters by company and subject, with cited regulations.
• Supplier Risk Report: Sanctions, VAT & Insolvency: one risk row per supplier from eight official registers, with changes since last run.
• Brand Impersonation Domain Monitor: Look-alikes: look-alike domains and certificates that imitate your brand, with a risk score.

Media and AI

• Podcast Mention Alerts: Brand Mentions by Second: where your brand is said in new podcast episodes, with the exact second.
• News Brand Mentions Monitor: New Articles Only: new news articles about your brand, people or topics.
• YouTube Channel Monitor: New Videos & Growth: new videos and view growth of the channels you watch.

Market signals

• US WARN Layoff Notices: 12 States Daily Feed: layoff and plant closure notices from official state sources.
• US Product Recalls Monitor: FDA & CPSC Feed: FDA and CPSC recalls in one feed, with severity.
• EU Tenders Feed: TED Public Procurement Alerts: EU public tenders from TED with CPV, country and value filters.
• Liquor License Openings: New Bars & Stores (US): new US liquor licences from official state open data.
• Clinical Trials Change Monitor: Status & Updates: status, enrollment and results changes on ClinicalTrials.gov studies.
• Career Site Jobs Monitor: ATS Jobs & New Roles: open roles from company career sites, with new and closed roles since the last run.
• App Reviews Monitor: App Store & Google Play: app reviews from both stores, only new ones in monitor mode.
• US Federal Contracts Feed: SAM.gov & USAspending: SAM.gov contract opportunities with the buyer's award history.
• Grants.gov Opportunities: US Federal Grants Feed: US federal grant opportunities by keyword, eligibility and agency.
• Federal Register Monitor: US Rules & Notices Feed: new US rules, proposed rules and notices by agency or topic.
• B2B Sales Trigger Feed: Hiring + Tech Leads: companies hiring for a role and using (or not using) a technology, with contacts.
• Local Business Leads: OpenStreetMap, No Website: local businesses by city and category, including those with no website.
• Solana Token Pair Monitor: New & Trending Pairs: new and trending Solana pairs with liquidity, volume and safety flags.

Feedback

Missing a field you want to track? Open an issue on the Issues tab.