🎬 Instagram Transcript Scraper & Reels Monitor

Turn public Instagram Reels into searchable transcripts, subtitles, hooks, keyword mentions, and profile-level insights — without an Instagram login.

Use one Actor for three jobs: transcribe specific Reel URLs, research the latest Reels from many public profiles, or monitor profiles and return only newly published videos.

✨ Why use this Actor?

• 📝 Get a full spoken-word transcript from every Reel with detectable speech
• ⏱️ Receive timestamped segments plus ready-to-use SRT and WebVTT subtitles
• 🎯 Detect opening hooks, spoken brands, keywords, calls to action, offers, discounts, and promo codes
• 👤 Research many public profiles in one run and receive a clean summary for each profile
• 🔔 Monitor competitors or creators and process only new Reels on later scheduled runs
• ⚡ Use genuine Instagram timed captions when available, with automatic speech-to-text fallback
• 📦 Export results as JSON, CSV, Excel, XML, RSS, SRT, or VTT
• 🔌 Connect the output to Make, Zapier, webhooks, schedules, or your own API

No cookies or Instagram credentials are required. The Actor works with public profiles and public /reel/, /p/, and /tv/ URLs.

🚀 Quick start

1. Open the Input tab.
2. Choose a workflow: Reel URLs, Profile research, or Monitor new Reels.
3. Add public Reel URLs or Instagram usernames.
4. Keep Fastest selected unless you specifically need a larger speech model.
5. Click Start and open the Reels & transcripts Dataset when the run finishes.

The Actor runs with a tested 4 GB memory configuration automatically. You do not need to tune memory or concurrency.

🧭 Three modes

Mode	Best for	What it returns
Reel URLs	Transcribing known posts	One clean Reel record for each supplied URL or shortcode
Profile research	Competitor, creator, and brand research	Recent Reel records plus a separate profile summary for every profile
Monitor new Reels	Scheduled monitoring	Only unseen Reels after the initial baseline, plus updated profile summaries

🔗 Reel URLs

Paste one or many public Instagram Reel, post, or IGTV links. Bare shortcodes are also accepted.

👥 Profile research

Add usernames such as nasa, @natgeo, or full profile URLs. Choose between 1 and 12 recent Reels per profile. One Reel per profile is the fastest setting for large lists.

Individual Reels stay in the Reels & transcripts Dataset. Aggregate rows are stored separately in Profile summaries, so exports remain clean and easy to use.

🔔 Monitor new Reels

Give the monitor a stable monitorId and reuse it on every scheduled run. The Actor remembers previously seen shortcodes in persistent storage.

For the first run, choose:

• Transcribe current Reels — process current posts immediately and create the baseline.
• Baseline only — remember current posts without transcribing them; later runs return only newer Reels.

🎛️ Input

Example: research multiple profiles

{
  "mode": "profiles",
  "profiles": ["nasa", "natgeo", "openai"],
  "maxReelsPerProfile": 1,
  "spokenKeywords": ["AI", "space", "new"],
  "hookSeconds": 5,
  "language": "auto",
  "transcriptionModel": "tiny",
  "preferNativeCaptions": true
}

Main input options

Field	Meaning	Default
mode	urls, profiles, or monitor	profiles
urls	Public Reel/post URLs or shortcodes	Example URL
profiles	Public usernames, @usernames, or profile URLs	nasa
maxReelsPerProfile	Recent Reels inspected per profile, from 1 to 12	1
monitorId	Stable identifier used to remember seen Reels	Generated from profiles
firstRunMode	processCurrent or baselineOnly	processCurrent
spokenKeywords	Spoken words or brand names to locate with timestamps	Empty
hookSeconds	Length of the opening hook, from 1 to 15 seconds	5
language	Automatic detection or a forced language code	auto
transcriptionModel	tiny, base, or small	tiny
preferNativeCaptions	Use timed Instagram captions before speech-to-text	true

📊 Output

Reels & transcripts

Each Reel record contains:

• Reel URL, shortcode, creator, coauthors, caption, publication date, cover, and media URL
• Views, likes, comments, duration, video dimensions, and audio status
• transcriptStatus: transcribed or no_speech
• transcriptionSource: instagram_native_captions, speech_to_text, or media_metadata
• Detected language and confidence
• Full transcript and timestamped segment JSON
• SRT and WebVTT subtitle text
• Opening hookText
• Matched keywords with timestamped context
• Calls to action, offer mentions, and promo codes
• sourceProfile and isNew for monitoring workflows

Profile summaries

The separate Profile summaries Dataset contains profile name, biography, follower counts, verification, discovered and processed Reel counts, frequent hooks, matched keywords, offers, and the top Reel by views.

Issues and run summary

• Recoverable problems are written to the separate Issues Dataset.
• A machine-readable run summary is saved as the OUTPUT record in the default key-value store.
• A run can complete as SUCCEEDED_WITH_ISSUES in OUTPUT while preserving every successful result.

🧾 Log

The log is compact and automation-friendly. It shows the selected mode, progress for each target, recoverable errors, and a final counter summary.

INFO  START mode=profiles urls=0 profiles=2 maxReels=1 model=tiny preferNativeCaptions=True requestConcurrency=2 jobConcurrency=2 transcriptionConcurrency=1
INFO  PROFILE_DONE profile=nasa discovered=1 processed=1 transcribed=1 noSpeech=0
INFO  PROFILE_DONE profile=natgeo discovered=1 processed=1 transcribed=1 noSpeech=0
INFO  FINISH status=SUCCEEDED mode=profiles reelsDiscovered=2 reelsSaved=2 reelsTranscribed=2 reelsNoSpeech=0 profilesProcessed=2 profileSummaries=2 issues=0 durationSeconds=18.42

Warnings such as URL_ERROR, PROFILE_ERROR, or TRANSCRIPT_ERROR include the affected input and are also recorded in the Issues Dataset.

💳 Pricing behavior

When pay-per-event pricing is enabled, a successful spoken-word transcription is the billable event. Reels with no detectable speech and failed inputs are not charged as successful transcriptions. Platform resource usage may be shown separately in the Actor's Pricing tab.

⚠️ Limits

• Only public profiles and public posts are supported.
• Instagram currently exposes up to 12 recent Reels per profile through the public profile feed.
• Direct URL mode can process older public Reels when you provide their links.
• Instagram can change or remove public media at any time.
• Runtime depends on video duration, speech density, the selected model, and Instagram response times.
• Collaboration posts can appear on one profile while retaining another primary author. Use sourceProfile and authorUsername to distinguish them.

❓ FAQ

Does it require an Instagram account or cookies?
No. It uses publicly available pages and media.

Does it support private profiles?
No. Private or login-gated media cannot be processed.

Why is a transcript empty?
Check transcriptStatus. Music-only videos, silent clips, or videos without a usable audio stream are returned as no_speech instead of being presented as failed transcriptions.

Where are the profile summaries?
Open the separate Profile summaries Dataset. They are intentionally kept out of the Reel table.

Can I schedule monitoring?
Yes. Create a task with mode: "monitor", keep the same monitorId, and attach an Apify schedule.

Can I export subtitles?
Yes. Use the subtitle Dataset view or read transcriptSrt and transcriptVtt from each Reel record.