Loom Scraper: Transcripts, Subtitles & Video Downloader avatar

Loom Scraper: Transcripts, Subtitles & Video Downloader

Pricing

from $2.40 / 1,000 video extracteds

Go to Apify Store
Loom Scraper: Transcripts, Subtitles & Video Downloader

Loom Scraper: Transcripts, Subtitles & Video Downloader

Extract public Loom video transcripts with timestamps, SRT/VTT subtitles, metadata (owner, views, duration) and MP4 or HLS download URLs. Paste share links, embed URLs or video IDs in bulk. Pure HTTP, no login or API key.

Pricing

from $2.40 / 1,000 video extracteds

Rating

5.0

(1)

Developer

Automation Lab

Automation Lab

Maintained by Community

Actor stats

2

Bookmarked

173

Total users

19

Monthly active users

17 hours ago

Last modified

Features

Limited permissionsMulti-tenantServer interface

Share

Loom Scraper

Extract transcripts, SRT/VTT subtitles, metadata, and video download URLs from Loom videos — no API key, no Loom account, no browser needed. Paste a public Loom share link, embed/iframe source URL, or raw video ID and get structured data in seconds. Accepted inputs are normalized to canonical share URLs and deduplicated by video ID.

What does Loom Scraper do?

Loom Scraper is an all-in-one data extraction tool for Loom, the async video messaging platform used by millions of B2B teams. It accesses Loom's public GraphQL API directly (pure HTTP — no Playwright, no headless browser) to extract:

  • 📝 Full transcripts with per-segment timestamps (startSeconds, endSeconds, text) and full concatenated plain text for LLM pipelines
  • 💬 Ready-to-use subtitles as WebVTT (transcriptVtt) and SRT (transcriptSrt) text for video, archive, and documentation tools
  • 📊 Video metadata — title, description, duration, owner, view count, comment count, emoji count, thumbnail URL
  • 🎬 Video download URLs — a signed MP4 link when Loom has one, plus an HLS (.m3u8) stream link that also covers newer videos with no MP4 file (signed, download promptly)
  • 🔗 Accepts public Loom /share/ links, /embed/ iframe source URLs, and raw 32-character video IDs
  • 🔒 Public Loom videos only — private, password-protected, and login-required videos are not supported

Batch-process hundreds of Loom videos in a single run and receive metadata, transcripts, subtitle formats, and download URLs together in one structured result.

Who is it for?

📚 Documentation & Knowledge Teams

Content teams who receive Loom walkthroughs from engineers, designers, or sales reps and need to convert video narration into written docs, wikis, or SOPs. Instead of watching hours of video, extract all transcripts in bulk and feed them into your documentation workflow.

🤖 AI / LLM Developers

Builders creating RAG pipelines, AI meeting summarizers, or video-to-document converters. Loom Scraper gives you clean, timestamped transcript text ready to pass to GPT-4, Claude, or Gemini without manual copy-paste.

📋 Compliance & Archiving Teams

Legal, HR, and compliance teams that need to archive video communications for audit trails, e-discovery, or regulatory requirements. Extract and store transcript text and metadata for all shared Looms in your organization.

🔁 Workflow Automation Engineers

Engineers building CI/CD workflows that automatically process new Loom videos — trigger on new video URL → extract transcript → post to Slack, create Jira ticket, or update Confluence page.

📈 Sales & Marketing Analysts

Teams analyzing sales demo recordings or marketing explainer videos to extract talking points, review competitor content, or build training datasets from existing video libraries.

Why use Loom Scraper?

  • ✅ All-in-one workflow — transcript segments, WebVTT/SRT subtitles, metadata, and download URL in a single run
  • ✅ Pure HTTP — no browser overhead, runs in 256 MB, blazing fast
  • ✅ Flexible public Loom inputs — use share links, embed/iframe source URLs, or a raw video ID
  • ✅ No API key required — works with public Loom videos
  • ✅ Bulk processing — process hundreds of videos in one run, with duplicate IDs processed once
  • ✅ LLM-ready output — transcriptText field gives plain text ready for AI pipelines
  • ✅ Timestamped segments — transcript array has startSeconds/endSeconds/text for each caption block
  • ✅ Apify platform — schedule runs, export to CSV/JSON/Excel, connect to 5,000+ apps via Zapier/Make
  • ✅ API access — integrate into any workflow via REST API, Node.js, or Python SDK

What data can you extract?

FieldDescription
idLegacy Loom video ID field (preserved for compatibility)
videoIdLoom video ID (32-character hexadecimal value)
urlCanonical public Loom share URL
titleVideo title
descriptionVideo description (if set)
privacyVisibility: public, private, etc.
durationSecondsDuration in seconds
durationMsDuration in milliseconds
createdAtISO 8601 creation timestamp
updatedAtISO 8601 last-updated timestamp
ownerIdOwner's Loom user ID
ownerNameOwner's display name
ownerEmailOwner's email (if publicly visible)
thumbnailUrlThumbnail image URL
viewCountTotal view count
distinctViewCountUnique viewer count
emojiCountTotal emoji reactions
commentCountTotal comment count
widthVideo width in pixels
heightVideo height in pixels
videoUrlSigned MP4 direct download URL, or null when Loom keeps no MP4 file for the video
videoStreamUrlSigned HLS (.m3u8) stream URL; use it with ffmpeg or yt-dlp when videoUrl is null
transcriptStatussuccess, unavailable, or null
transcriptArray of {startSeconds, endSeconds, text} segments
transcriptTextFull transcript as a single plain-text string
transcriptVttReady-to-save WebVTT subtitle text, or null when unavailable/not requested
transcriptSrtReady-to-save SRT subtitle text, or null when unavailable/not requested
errorExtraction error message; null for successful records
scrapedAtISO 8601 timestamp of extraction

How much does it cost to scrape Loom videos?

This actor uses pay-per-event (PPE) pricing — you pay only for what you extract.

TierPer video (video-extracted)Actor start (flat, once per run)
FREE$0.003$0.002
BRONZE$0.003$0.002
SILVER$0.0027$0.002
GOLD$0.0024$0.002
PLATINUM$0.0024$0.002
DIAMOND$0.0024$0.002

The start fee is charged only when the run saves its first video, so a run that finds only deleted or private videos costs nothing.

Real-world cost examples (FREE tier, $0.003/video + $0.002/run):

TaskVideosRunsApprox. Cost
Quick transcript check11~$0.005
Weekly team video archive201~$0.062
Bulk knowledge base build1001~$0.302

With $5 free credits, you can extract roughly 1,650 videos in batch runs of 100, or about 1,000 videos if running one video per run.

How to scrape Loom videos

  1. Go to the Loom Scraper page on Apify Store
  2. Click Try for free
  3. In the Videos section, add one or more public Loom share links, embed/iframe source URLs, or raw 32-character video IDs
  4. Choose what to extract: transcript, metadata, and/or video download URL
  5. Click Save & Run
  6. Download results as JSON, CSV, or Excel — or connect directly to your app via API

Example input:

{
"videoUrls": [
"https://www.loom.com/share/002508e86fde4232bb8de474eb5c65c4",
"https://www.loom.com/embed/0054b03fe58a40ea960ef2e48a333667",
"002508e86fde4232bb8de474eb5c65c4"
],
"extractTranscript": true,
"extractMetadata": true,
"extractVideoUrl": true,
"includeFullTranscriptText": true
}

Metadata-only mode (fastest):

{
"videoUrls": ["https://www.loom.com/share/002508e86fde4232bb8de474eb5c65c4"],
"extractTranscript": false,
"extractMetadata": true,
"extractVideoUrl": false
}

Input parameters

ParameterTypeDefaultDescription
videoUrlsArrayrequiredPublic Loom share links, embed/iframe source URLs, or raw video IDs. Normalized and deduplicated by ID
extractTranscriptBooleantrueExtract timed transcript segments
extractMetadataBooleantrueExtract title, owner, views, duration, etc.
extractVideoUrlBooleantrueExtract signed MP4 and HLS stream URLs
includeFullTranscriptTextBooleantrueInclude full transcript as plain text string

Output examples

Single video with full transcript:

{
"id": "002508e86fde4232bb8de474eb5c65c4",
"videoId": "002508e86fde4232bb8de474eb5c65c4",
"url": "https://www.loom.com/share/002508e86fde4232bb8de474eb5c65c4",
"title": "Screen and engagement tips",
"description": null,
"privacy": "public",
"durationSeconds": 154,
"durationMs": 154000,
"createdAt": "2022-08-05T21:16:35.487Z",
"ownerName": "Miguel Acero",
"ownerEmail": null,
"thumbnailUrl": "https://cdn.loom.com/sessions/thumbnails/002508e86fde4232bb8de474eb5c65c4-1659734506699.jpg",
"viewCount": 94895,
"distinctViewCount": 88557,
"emojiCount": 5144,
"commentCount": 477,
"width": 1152,
"height": 720,
"videoUrl": "https://cdn.loom.com/sessions/transcoded/002508e86fde...mp4?Policy=...",
"videoStreamUrl": "https://luna.loom.com/id/002508e86fde.../resource/hls/playlist-multibitrate.m3u8?Policy=...",
"transcriptStatus": "success",
"transcript": [
{ "startSeconds": 0.31, "endSeconds": 4.56, "text": "Hey, Loom community..." },
{ "startSeconds": 4.71, "endSeconds": 9.20, "text": "that supports agents..." }
],
"transcriptText": "Hey, Loom community... that supports agents...",
"transcriptVtt": "WEBVTT\n\n00:00:00.310 --> 00:00:04.558\nHey, Loom community...\n",
"transcriptSrt": "1\n00:00:00,310 --> 00:00:04,558\nHey, Loom community...\n",
"error": null,
"scrapedAt": "2026-03-31T03:00:00.000Z"
}

Tips for best results

  • 🔗 Use a supported public Loom input — /share/VIDEO_ID, /embed/VIDEO_ID (the URL from an iframe's src), or the raw 32-character video ID
  • 🧹 Duplicates are automatic — repeated share, embed, and raw-ID forms of the same video are processed once, in first-seen order
  • ⚡ Start small — test with 1-2 inputs first to verify output before bulk runs
  • 📥 Download video URLs promptly — signed MP4 and stream URLs expire within a few hours
  • 🎞️ No MP4? — newer Loom videos are stored only as streams; download them with ffmpeg -i "<videoStreamUrl>" -c copy video.mp4
  • 🧾 Empty values are null — a value Loom does not publish (such as the owner's email) is null, never an empty string
  • 💾 Transcript-only mode — if you only need transcripts, disable extractVideoUrl for slightly faster runs
  • 🤖 LLM integration — use transcriptText for direct GPT/Claude input; use transcript when you need structured timestamps
  • 💬 Subtitle export — save transcriptVtt as .vtt or transcriptSrt as .srt; both are produced from the same caption fetch with no extra request
  • 📅 Schedule recurring runs — set up daily runs to archive new Looms automatically via Apify scheduling
  • 🔒 Private videos — this actor only works on public Loom videos. Private, password-protected and deleted videos are skipped, are not charged, and are counted in the run's status message

Integrations

🗂️ Loom → Google Sheets (documentation pipeline)

Connect Loom Scraper to Google Sheets via Zapier. Every time you run the scraper, transcript text and metadata are automatically added as new rows — building a searchable video archive without manual copy-paste.

🤖 Loom → OpenAI/Claude (AI summarization)

Feed transcriptText output directly into a Make.com or n8n scenario that calls GPT-4 or Claude API. Generate summaries, action items, or key points from Loom walkthroughs automatically.

📬 Loom → Slack (team digest)

Schedule a daily Apify run on a batch of Loom URLs, then post a digest of titles + transcript excerpts to your team's Slack channel. Great for async standups and knowledge sharing.

🗄️ Loom → Notion/Confluence (knowledge base)

Use webhooks: when a Loom run completes, a webhook triggers a Make scenario that creates a Notion page or Confluence article with the transcript, title, and thumbnail — turning every Loom into a searchable text document.

♻️ Scheduled archiving workflow

Set up a scheduled Apify run (daily or weekly) with a list of Loom share URLs. Combined with Apify's dataset export, this builds an automatic rolling archive of all your shared Looms.

API usage

Node.js

import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: 'YOUR_API_TOKEN' });
const run = await client.actor('automation-lab/loom-scraper').call({
videoUrls: [
'https://www.loom.com/share/002508e86fde4232bb8de474eb5c65c4',
],
extractTranscript: true,
extractMetadata: true,
extractVideoUrl: true,
includeFullTranscriptText: true,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items[0].transcriptText);

Python

from apify_client import ApifyClient
client = ApifyClient(token="YOUR_API_TOKEN")
run = client.actor("automation-lab/loom-scraper").call(run_input={
"videoUrls": [
"https://www.loom.com/share/002508e86fde4232bb8de474eb5c65c4",
],
"extractTranscript": True,
"extractMetadata": True,
"extractVideoUrl": True,
"includeFullTranscriptText": True,
})
items = list(client.dataset(run["defaultDatasetId"]).iterate_items())
print(items[0]["transcriptText"])

cURL

curl -X POST https://api.apify.com/v2/acts/automation-lab~loom-scraper/runs \
-H "Authorization: Bearer YOUR_API_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"videoUrls": ["https://www.loom.com/share/002508e86fde4232bb8de474eb5c65c4"],
"extractTranscript": true,
"extractMetadata": true,
"extractVideoUrl": true
}'

Use with AI agents via MCP

Loom Scraper is available as a tool for AI assistants that support the Model Context Protocol (MCP).

Add the Apify MCP server to your AI client — this gives you access to all Apify actors, including this one:

Setup for Claude Code

$claude mcp add --transport http loom-scraper "https://mcp.apify.com/?tools=automation-lab/loom-scraper"

Setup for Claude Desktop, Cursor, or VS Code

Add this to your MCP config file:

{
"mcpServers": {
"apify": {
"url": "https://mcp.apify.com/?tools=automation-lab/loom-scraper"
}
}
}

Your AI assistant will use OAuth to authenticate with your Apify account on first use.

Example prompts

Once connected, try asking your AI assistant:

  • "Use automation-lab/loom-scraper to extract the transcript from https://www.loom.com/share/002508e86fde4232bb8de474eb5c65c4 and summarize the key points"
  • "Get all the metadata from this Loom video and tell me how many people watched it"
  • "Extract the transcripts from these 5 Loom videos and compile them into a single document"

Learn more in the Apify MCP documentation.

Legality

Loom Scraper only accesses publicly shared videos — videos that the owner explicitly made available via a share link. It does not bypass any authentication, access private workspaces, or circumvent any access controls.

Web scraping of publicly accessible content is generally legal in most jurisdictions. The hiQ v. LinkedIn ruling and similar cases have established that scraping publicly available data does not violate the Computer Fraud and Abuse Act (CFAA).

Always ensure you have permission to collect and use the video content and transcripts you extract. Respect intellectual property rights. This actor is intended for legitimate business use cases such as documentation, archiving, and AI pipeline development.

FAQ

How fast is the scraper? Very fast — pure HTTP, no browser. Each video takes 2-5 seconds to process (metadata + transcript). A batch of 100 videos typically completes in 3-5 minutes.

How much does it cost per video? On the FREE tier: $0.003 per video plus a $0.002 start fee per run, so a single-video run costs about $0.005. Higher plans pay less per video — $0.0024 on GOLD and above. See the pricing section above for full tier details.

Does it work without a Loom account? Yes. No account or API key required. The actor uses Loom's public GraphQL API which accepts anonymous requests for public videos.

How fast is one video? The actor uses pure HTTP, no browser, and runs in 256 MB. A video usually takes 1-3 seconds (metadata, transcript and download links together).

Why are some transcripts missing? Some Loom videos do not have transcripts — either because the owner disabled transcription, the video was recorded before Loom added AI transcription, or the transcription failed during processing. The transcriptStatus field will show unavailable in these cases. Metadata (title, duration, views, etc.) is still extracted.

Why is the videoUrl empty for some videos? Newer Loom videos are stored only as streams, with no MP4 file, so videoUrl is null. Use videoStreamUrl instead: it is a signed HLS playlist that ffmpeg or yt-dlp turns into an MP4. Signed URLs are generated fresh each run and expire within hours.

Which Loom input formats can I use? Use a public https://www.loom.com/share/VIDEO_ID link, a https://www.loom.com/embed/VIDEO_ID iframe source URL, or the raw 32-character hexadecimal video ID. The actor normalizes each accepted input to a canonical share URL. Full iframe HTML, unrelated hosts, malformed IDs, and private/login URLs are rejected.

My video input returned no results — what's wrong? Read the run's status message: it says how many videos were saved and how many were not found, not public, or not valid Loom inputs. Videos must be publicly shared (not password-protected or private), and deleted videos return nothing. Skipped videos are never charged, and a run where every input is invalid fails at once without any charge.

Is my data used for anything else? When a run or request fails, the actor sends sanitized diagnostic input, the error, and the actor/build/run IDs to our private GlitchTip error tracker so we can fix it. Secret fields and URL query strings are removed, and reports are kept for 30 days.

Server API: one video per HTTP request

Loom Scraper is also a Server Actor. Besides normal runs, it answers short HTTP requests directly, without starting a run for each one: useful for AI agents and apps that need one transcript right now. Copy the server URL from the Actor's Server tab (do not build it from the Actor name) and authenticate with your Apify token. Normal Actor runs keep the full input and dataset output; use them for batches.

curl -sS "$SERVER_URL/video?url=https://www.loom.com/share/002508e86fde4232bb8de474eb5c65c4" -H "Authorization: Bearer $APIFY_TOKEN"
curl -sS "$SERVER_URL/video" -H "Authorization: Bearer $APIFY_TOKEN" -H 'Content-Type: application/json' -d '{"url":"002508e86fde4232bb8de474eb5c65c4","videoUrl":false}'
Method and pathInputOutput and charge
GET /Noneready: true and the routes; no scraping or charge
GET /videoQuery url (public Loom share URL, embed URL or 32-character video ID); optional transcript, videoUrl, transcriptText (true/false, default true){ok,count,results} with one record, the same fields as a dataset item; one video-extracted event, charged only when the video is returned; no start fee
POST /videoJSON body with the same fields: url, optional transcript, videoUrl, transcriptTextSame as GET /video

Bodies must be JSON and at most 4096 bytes. Each request has a 60-second deadline and a 2 MiB response cap, and one server works on at most four requests at once. Nothing is cached between callers. Errors use { "ok": false, "error": { "code": "...", "message": "...", "retryable": false } }: 400 invalid input or JSON, 402 spending limit, 404 video not found or not public (VIDEO_UNAVAILABLE, not charged) or unknown route, 405 wrong method, 413 size limit, 415 media type, 502 Loom or charging failure, 503 server busy or stopping, 504 timeout. Retry only when retryable is true. A successful answer can carry warning: { code: "CHARGE_INCOMPLETE" } when its charge did not go through; you are charged only for what went through. The Server tab shows the full OpenAPI description.