Loom Transcript Scraper - Captions to Text

Extract the captions that Loom already exposes for public share and embed videos. Give the actor public Loom links and receive clean LLM-ready text, timestamped segments, and optional VTT/SRT subtitle strings. It is transcript-first: it does not download video files, bypass passwords, or access private Loom accounts.

When to use it

Use it to turn public Loom walkthroughs, demos, training videos, or support recordings into text for RAG, documentation, summaries, accessibility captions, or content research. Do not use it for private, password-protected, folder, account, or video-download workflows.

Output per video

Field	Description
transcriptText	Full clean transcript, ready for prompts and embeddings.
transcriptSegments	Timestamped { startSeconds, endSeconds, text } caption units.
transcriptVtt, transcriptSrt	Optional subtitle strings derived from the timestamped transcript.
title, ownerName, durationSeconds	Public video context when metadata is enabled.
status, error	Honest result state; videos without captions are retained but never billed.

Pricing

Pay per event + platform usage is enabled.

Event	Price
Actor start	$0.00005
Public Loom transcript saved	$0.01/video

Videos with no transcript and upstream failures are not charged the per-video event. A run of 25 successful videos costs at most $0.25005 in actor events, plus Apify platform usage.

Input

{
    "loomUrls": [
        "https://www.loom.com/share/002508e86fde4232bb8de474eb5c65c4",
        "https://www.loom.com/embed/002508e86fde4232bb8de474eb5c65c4"
    ],
    "maxVideos": 25,
    "includeMetadata": true,
    "includeCaptionFiles": true
}

You may also add startUrls request objects in a programmatic workflow. The actor accepts share URLs, embed URLs, and raw 32-character Loom IDs, and deduplicates all supplied targets automatically.

Example output

{
    "videoId": "002508e86fde4232bb8de474eb5c65c4",
    "sourceUrl": "https://www.loom.com/share/002508e86fde4232bb8de474eb5c65c4",
    "status": "transcript_extracted",
    "title": "Product walkthrough",
    "durationSeconds": 154,
    "transcriptText": "Hello from Loom.\nThis is the second caption.",
    "transcriptSegments": [{ "startSeconds": 0.31, "endSeconds": 8.5, "text": "Hello from Loom." }],
    "transcriptVtt": "WEBVTT\n\n00:00:00.310 --> 00:00:08.500\nHello from Loom.",
    "transcriptSrt": "1\n00:00:00,310 --> 00:00:08,500\nHello from Loom.",
    "scrapedAt": "2026-07-15T10:00:00.000Z"
}

How it works

1. The actor validates a public Loom share or embed URL and normalizes it to the video ID.
2. It loads the public share page over HTTP and reads the page's embedded public video state.
3. It downloads Loom's short-lived signed transcript JSON and normalizes caption phrases.
4. It writes the usable transcript row and charges the per-video event in the same SDK call.

This approach does not use browser automation, third-party transcription, cookies, or API keys. Signed caption links are not returned because they expire quickly.

Run outcomes

OUTPUT and RUN_SUMMARY are written for every terminal path. Outcomes are COMPLETE, PARTIAL, VALID_EMPTY, INVALID_INPUT, or UPSTREAM_FAILED. A public video without captions becomes a non-billable no_transcript dataset row; a fully unavailable upstream becomes an honest failed run only when no useful transcript was produced.

MCP usage

Use this when an agent needs text from a known public Loom video. It returns one dataset row per input video; use transcriptText for concise retrieval and transcriptSegments for timestamp-aware answers. It is not a general Loom search, downloader, or private-account connector.

Limits and responsible use

• Only public, non-password-protected Loom videos are supported.
• The actor extracts captions Loom has made available; it does not generate speech-to-text for videos without captions.
• Loom can change its public page data or signed-URL behavior. Failures are surfaced in status and the run summary.
• You are responsible for having a lawful basis to process the public content and for complying with Loom's terms and applicable privacy laws.