Instagram Transcript Scraper: Reels and Videos to Text

Turn Instagram Reels and videos into text, without an account or API key. Paste reel or post links, or give an account, and each video comes back with its transcript as plain text and as timestamped segments, its language, optional SRT and WebVTT subtitle files, and the post's metadata (account, date, likes, comments, caption, sound).

Two things you would otherwise check by hand: reels that are only music come back as a free no_speech record instead of a made-up transcript, and every video that cannot be transcribed (deleted, age-restricted, photo post) gets a free record that says why. You pay only for transcripts delivered.

What we measured

On the Apify platform on 8 October 2026, with the default settings, on 43 links taken from Reddit posts that shared Instagram reels, plus two accounts:

• 25 transcripts in 6 languages (English, Spanish, French, Russian, Portuguese, Hindi), average length 59 seconds.
• 17 reels without clear speech (music, sounds, no audio track, loops such as "Peppino, Peppino, Peppino", one uncertain language) came back as free missing records: none was charged.
• 5 links were deleted posts or age-restricted reels: free error records with the reason.
• 47 records in 2 minutes 30. The default input (one reel) runs in about 6 seconds.

Running the same input again (17 records: Hindi reels, two Indian news accounts, unknown accounts, damaged links) gave the same verdict on every record, and Hindi came back in Hindi (Devanagari) every time (5 of 5 on the hardest reel).

Speech-to-text runs on third-party AI providers: OpenAI's Whisper models (large-v3 and large-v3 turbo) run by Groq for most videos, and ElevenLabs Scribe, a stronger model, to settle the language when hints disagree and as a backup. The audio is sent to them only to be transcribed. ElevenLabs stores each transcript on its side: we delete it as soon as we have read it and check that it is gone (we read it again and get "not found"); if a deletion cannot be confirmed before the run ends, the run log says so.

Whisper is known to invent short phrases on music ("Thank you.", "Music", "Bye.", subtitle credits). We measured that on 170 reels and TikTok videos and filter it: a phrase the model is not confident about, a known phantom phrase, or a few words spread over a long stretch of music are not speech. A video whose speech is all filtered out is missing with missingReason: "no_speech", free. We never sell part of a video as its transcript: when the filter has to leave out more than 20% of the speech (the model looped or was unsure over that part), the video is missing with missingReason: "incomplete_transcript", free. The same holds when Whisper hits its own limit: its output is capped for each 30 seconds of audio, and dense speech in some scripts (Gujarati, Telugu, fast Hindi) needs more, so the end of each 30 seconds is silently dropped. We detect it on every window; such a reel goes to a second, stronger speech-to-text model, and if no model writes it in full, it is a free incomplete_transcript, never a charged fragment. Loops are measured per character, so Hindi, Urdu, Arabic or Tamil text is judged like English (until 8 October 2026 a byte-based measure dropped correct Hindi sentences as "loops").

What you can transcribe

Input	Examples	What you get
Reels	https://www.instagram.com/reel/DbIdlf0j6uA/, /reels/…, /{account}/reel/…	that reel
Video posts	https://www.instagram.com/p/…, /tv/…, instagr.am/p/…, share links (/share/…)	the video; for a carousel, each of its videos
Accounts	@natgeo, https://www.instagram.com/natgeo/, …/natgeo/reels/	the account's latest videos

Mix them in one run. A video found in two inputs (a link and an account) is transcribed and charged once.

Accounts: without signing in, Instagram shows the 12 latest posts of an account (pinned posts first) and no more; we read those, skip the photo posts, and transcribe up to maxVideosPerAccount videos. For older videos, paste their links.

Carousels: each video of a carousel is its own record (carouselIndex 1, 2…), charged as one transcript each. Photo carousels do not take the place of a video: the account's list is read on until maxVideosPerAccount videos are found, and if the 12 posts Instagram shows do not hold that many, the run says so in its status.

Languages

Whisper's fast model sometimes mistakes one language for another; on Hindi reels it then writes an invented English "translation", or Hindi in the Urdu script. We do not trust a single detection when anything says otherwise:

1. Hints: the caption written by the author (its script: Devanagari, Hangul, Japanese kana…, or Hindi written in Latin letters, "Hinglish": "ke", "se", "mein", "nahi", "lekar"…). English is not used as a hint (Whisper only ever translates into English, so an English hint cannot catch a translation).
2. When a hint and the detection disagree, a stronger speech-to-text model settles the language and transcribes the reel. Hindi and Urdu reels (by their caption) go to it directly, and so does any reel the fast model hears as Hindi, Urdu or Punjabi without a hint (it mixes the three up). On 120 replayed decisions (8 October 2026) the ElevenLabs referee made no wrong call; when it is not available for a reel, Whisper large-v3 settles the language and must agree, or nothing is sold. The referee is paid per use, so a run calls it only once the transcripts it has delivered cover the cost: in a larger run such a reel waits for the others; sent alone, or in a run that delivers little, it can come back as a free uncertain_language record.
3. Hindi and Urdu are one spoken language in two scripts: the caption's script decides.
4. If the language still cannot be settled, the reel comes back as a free uncertain_language record, never as a charged guess, and never as a translation.

language is a preference, never an instruction: it changes neither the models used nor the number of passes, and never forces a translation. A reel in English with language: "fr" comes back in English with requestedLanguageFound: false. With language set to Hindi, Urdu or Punjabi, a reel that Whisper hears as English (with no caption saying otherwise) comes back as a free uncertain_language record: on those languages Whisper sometimes writes an English translation instead of what was said.

Output

One record per video. A transcript (status: "ok"), shortened here:

{
  "status": "ok",
  "platform": "instagram",
  "videoId": "DbIdlf0j6uA",
  "url": "https://www.instagram.com/reel/DbIdlf0j6uA/",
  "carouselIndex": null,
  "author": {
    "username": "natgeo",
    "fullName": "National Geographic",
    "id": "787132",
    "isVerified": true,
    "url": "https://www.instagram.com/natgeo/"
  },
  "caption": "Describe Pompeii: Out of Time in 3 words? It's Now Streaming. …",
  "hashtags": ["pompeiioutoftime"],
  "publishedAt": "2026-07-23T13:00:03Z",
  "isReel": true,
  "viewCount": null,
  "likeCount": 28343,
  "commentCount": 320,
  "sound": { "title": "Original audio", "author": "natgeo", "isOriginal": true },
  "durationSeconds": 36.2,
  "language": "en",
  "transcriptSource": "speech_to_text",
  "isAutoGenerated": true,
  "hasSpeech": true,
  "transcript": "Three words to describe this series about Pompeii. History. There's something you're going to want to see. …",
  "segments": [
    { "start": 0, "end": 4, "duration": 4, "text": "Three words to describe this series about Pompeii." },
    { "start": 4, "end": 5, "duration": 1, "text": "History." }
  ],
  "segmentCount": 14,
  "wordCount": 70,
  "characterCount": 432,
  "billedMinutes": 1,
  "input": "https://www.instagram.com/reel/DbIdlf0j6uA/",
  "inputType": "video",
  "position": null,
  "scrapedAt": "2026-10-08T01:07:51.288Z"
}

A video without a transcript comes back as a free record that holds the reason and nothing else (no account, counts or caption: those come with transcripts):

{
  "status": "missing",
  "platform": "instagram",
  "videoId": "DeLb7gYJNiV",
  "url": "https://www.instagram.com/reel/DeLb7gYJNiV/",
  "missingReason": "no_speech",
  "message": "no speech detected (music, sounds or silence); free",
  "input": "https://www.instagram.com/reel/DeLb7gYJNiV/",
  "inputType": "video"
}

Fields that need a word of explanation

• status: ok (transcript, charged), missing (the video exists but has nothing to transcribe: no_speech, incomplete_transcript, no_audio, photo_post, or too_long past your limit; free) or error (the link could not be read; free).
• errorCode and retryable: not_found, age_restricted, invalid_url, account_not_found, account_private, account_empty are final; private (Instagram showed the post as private), blocked, network, transcription_capacity and transcription_failed are worth retrying later (retryable: true).
• language: ISO 639-1 code of the language spoken and transcribed. We never translate: if you set language and a video is in another language, you get the language actually spoken, with requestedLanguageFound: false.
• segments: start, end and duration in seconds. includeSrt and includeVtt add ready-to-save .srt and .vtt files.
• viewCount: Instagram hides play counts from visitors who are not signed in, so it is usually null.
• filteredSegments: segments of the raw speech-to-text output left out of the transcript as non-speech (a "Thank you." over silence, a "[Music]" tag, a model loop); minutes are billed on the video's length as usual. A model loop or an unsure passage that held more than 20% of the speech makes the video a free incomplete_transcript record instead.
• billedMinutes: the started minutes of video billed for this transcript (a 36-second reel is 1 minute).
• input and position: which of your inputs produced the record, and the rank of the video in an account's list. Records are written as each video finishes: sort by input and position for the source's order.

The dataset has three views: Overview, Transcripts and Video details. For a spreadsheet, export the Transcripts view as Excel (XLSX): one row per video, and Excel treats every cell as text. The full record with segments does not fit a spreadsheet (one column per segment), and a CSV opened in Excel may read a caption that starts with =, +, - or @ as a formula: open CSV files as text, or prefer XLSX or JSON. Excel also cuts a cell after 32,767 characters (about 35 minutes of speech). Control characters and text-direction marks are removed from captions and transcripts. The run's summary is the OUTPUT record of the default key-value store: per input, videos, transcripts, missing, errors and duplicates, plus speech-to-text and network statistics.

Input examples

The latest videos of two accounts and two reels, with subtitle files:

{
  "urls": ["@natgeo", "https://www.instagram.com/bbcnews/", "https://www.instagram.com/reel/DbIdlf0j6uA/"],
  "maxVideosPerAccount": 6,
  "includeSrt": true
}

Text only, Spanish reels, at most 200 transcripts:

{
  "urls": ["https://www.instagram.com/reel/…/", "https://www.instagram.com/reel/…/"],
  "language": "es",
  "transcriptFormat": "text",
  "maxItems": 200
}

Pricing

Pay per result: $0.001 per transcript plus $0.0045 per started minute of video transcribed, plus a start fee of $0.004 per GB of memory per run ($0.004 at the default 512 MB). A 40-second reel costs $0.0055, a 2-minute video $0.010. Videos without speech, incomplete transcripts, deleted or private posts, errors and duplicates are free. Platform usage is included.

Apify plan	Per transcript	Per started minute
Free	$0.001	$0.0045
Bronze	$0.001	$0.0040
Silver	$0.001	$0.0036
Gold and above	$0.001	$0.0034

Your maximum charge per run is never exceeded: the length of each video is read before it is transcribed, and the run stops at the last transcript it can charge and says so in its status. maxItems caps the transcripts of a run: ask for 10, get 10.

Free records are a by-product, not a product: a run stops, and says why, once 100 videos or inputs (an unknown account counts) plus 2 per transcript have given no transcript, or once 15 minutes of audio plus 2 per billed minute have been transcribed for nothing (music-only reels); audio we could not send because our capacity was used up does not count. maxDurationMinutes (default 30, at most 60) skips longer videos before anything is transcribed. A single video that would cost more than what is left of your maximum charge gets a free over_budget record and the run goes on with cheaper ones.

Use cases

• Content research: what creators in a niche say in their reels, searchable and quotable.
• Repurposing: turn your own reels into captions, blog posts, newsletters and subtitles.
• LLM pipelines: summaries, topic tagging, hooks and scripts analysis on hundreds of reels.
• Accessibility: SRT and VTT files to edit or upload.

Limits

• Speech or music. The speech-to-text service does not return exactly the same text on every pass over the same reel; on clear speech the difference does not matter, on music it can. So we keep a margin: a reel is charged only when its speech is clear (enough words, good confidence, no loops such as "wrestles, wrestles, wrestles"); everything in between comes back as a free no_speech record. Clearly sung lyrics still count as speech. Short or distant speech over loud music (a slogan shouted once) may be classed as no_speech.
• Accounts: the 12 latest posts only (Instagram's limit without an account). No stories, highlights, tagged posts, hashtags or search.
• Private accounts and age-restricted reels need a signed-in account; we do not use accounts, so they come back as free error records.
• Instagram does not show captions (subtitles) to visitors: every video goes through speech-to-text.
• Speed. Our speech-to-text capacity is shared fairly between runs and paced: count about 3 seconds per video on large runs. If it is temporarily used up, the videos that need it get free transcription_capacity records (retryable: true, "temporarily used up, retry later") and the status says when to retry; they never count against your run's limits and never make it fail.
• Longest video: 60 minutes (maxDurationMinutes, 30 by default).
• Connection: keep the default (automatic). If you set your own connection and it refuses everything, the run stops at once with a free proxy_failed record per input and says so.

Memory and speed

512 MB is the default and the best setting: the run waits on speech-to-text, not on memory, and the Actor uses under 100 MB. More memory does not make it faster (the maximum is 1 GB).

FAQ

Is it legal to scrape Instagram transcripts? This Actor reads only public posts, the way a visitor who is not signed in sees them, and accepts no Instagram terms on your behalf. In Meta v. Bright Data (N.D. Cal., January 2024) the court found that collecting public data while logged out did not breach Meta's terms; it is a first-instance decision, not binding everywhere. Transcripts reproduce what people say, which may be protected by copyright and may contain personal data: check that your use (research, analysis, quotation, accessibility) is allowed where you are, and handle personal data under the rules that apply to you (GDPR in Europe). We keep nothing after the run; the transcripts ElevenLabs stores are deleted during the run and the deletion is checked (the run log says so if one could not be confirmed).

Do I need an Instagram account or any setup? No. Keep the default connection: its cost is included in the price.

Why is a reel "no_speech" when I hear someone? Check OUTPUT.speechFilter for that video: very short or very uncertain speech over music is filtered. If a reel with clear speech comes back no_speech, open an issue with the link: we answer within 48 hours.

What happens when Instagram changes something? The Actor runs every day on known reels, so we see a change before you do. A page Instagram refuses is retried automatically, and a run that fails on our side ends as Failed and charges no transcript.