YouTube Shorts Transcripts

Video walkthrough

A 50-second recording of a real run (one channel, 8 Shorts, 20 comments each): https://youtu.be/0CwQeIuzTg0

Get the full spoken text of any YouTube Short, with timestamps, plus its title, channel, duration and language.

Paste a list of Shorts (or a whole channel) and get one clean record per Short: the transcript split into timestamped fragments and the plain text ready to use.

You pay per transcript delivered — Shorts without captions are reported as sin_subtitulo and are not charged.

What you get

One dataset item per Short:

Field	Description
video_id, url	Short identifier and link
titulo, canal	Video title and channel name
duracion_s	Duration in seconds
idioma	Language of the transcript found
subtitulo_automatico	true if it is YouTube's auto-generated caption, false if it is a human-made one
es_traduccion	true if the text is one of YouTube's automatic machine translations instead of the video's original language
texto_completo	The full transcript as one plain-text string
texto_con_tiempos	The same text, one line per fragment, prefixed with its time ([00:01.64] …) — ready for a spreadsheet
fragmentos_total	Number of fragments
estado, error	ok, or the reason the Short had no transcript

Input

• Start URLs — Shorts links, video links, or channel URLs (e.g. https://www.youtube.com/@MrBeast/shorts). Channel URLs return every Short shown on the channel's Shorts tab.
• Max Shorts — cap per run.
• Languages — preferred transcript languages, in order. A human-made caption is always preferred over an auto-generated one.
• Return the video's original language — on by default: you get the language the video was really recorded in, not one of YouTube's automatic machine translations. Turn it off only if you deliberately want translations to your preferred languages.
• Proxy — residential Apify Proxy by default. YouTube rate-limits transcripts per IP, so the tool rotates the IP on every request and retries on 429, 403 and YouTube's "sign in to confirm you're not a bot" response.

Use cases

• Content repurposing — turn a creator's Shorts into posts, newsletters or scripts.
• Research — analyse what is being said in short-form video at scale.
• Creator monitoring — follow competitors' Shorts and read the text without watching.
• AI / RAG pipelines — clean text with timestamps, ready to chunk and embed.
• Subtitles and editing — timestamps of every line, straight from the source.
• Accessibility and moderation review — searchable text for a batch of Shorts.

Output example

{
  "video_id": "aNI2jv7lBPQ",
  "url": "https://www.youtube.com/shorts/aNI2jv7lBPQ",
  "titulo": "Giving $10,000 To A Random Subscriber",
  "canal": "MrBeast",
  "duracion_s": 40,
  "idioma": "en",
  "subtitulo_automatico": true,
  "fragmentos_total": 35,
  "texto_completo": "Find a subscriber for whatever country that lands on, give them 10 grand. ...",
  "fragmentos": [{ "t": 0.0, "d": 4.2, "x": "Find a subscriber for whatever country" }],
  "estado": "ok"
}

Pricing

Pay per event: US$5.00 per 1,000 transcripts (US$0.005 per Short with a transcript). Shorts without captions are free.

Typical checks: this tool is tested against real channels. On a 40-Short channel in English it returned 38 transcripts (95%); the two missing ones had no captions at all. On a 12-Short channel in Spanish it returned 12 of 12. Runs are capped by Max Shorts.

Nota en español

Esta herramienta saca el texto completo de cualquier YouTube Short, con sus tiempos, más el título, el canal, la duración y el idioma. Le pasas enlaces de Shorts o un canal entero y devuelve una fila por Short, con los fragmentos marcados por tiempo y el texto listo para usar.

Solo se cobra por transcripción entregada: los Shorts que no traen subtítulo salen marcados como sin_subtitulo y no se cobran. Precio: US$5 por cada 1.000 transcripciones.

Sirve para reutilizar contenido, investigar, monitorear canales, alimentar flujos de IA o sacar subtítulos para edición. En las pruebas reales: 95% de Shorts resueltos en un canal en inglés y 100% en un canal en español.

Example script / Script de ejemplo

A ready-to-run example (Python, no dependencies required) that runs this Actor over a channel and writes the result to CSV lives here: https://github.com/sictegltda-cyber/youtube-shorts-transcripts-to-csv

Its README documents the three traps this Actor solves (machine-translated captions, the 1.9 MB watch page, and the per-IP rate limit), with the measurements.

---