TED Talks Scraper

Scrape TED.com talks — title, speaker, duration, view count, publish/record dates, topics, language, description, and thumbnail. Two modes: fetch specific talks by URL/slug, or browse all talks in a topic.

Pure HTTP, no auth, no proxy required. TED.com works from datacenter IPs.

What you get

Talk records (recordType=talk)

Field	Description
id	TED talk ID
slug	URL slug (e.g. sir_ken_robinson_do_schools_kill_creativity)
url	Canonical talk URL
title	Talk title
speaker	Presenter name (presenterDisplayName)
partnerName	TEDx / TED Conference / Independent / etc.
description	Plain-text summary
socialDescription	Optional alt summary used on social embeds (only when different)
durationSeconds	Talk length in seconds
durationFormatted	Human-readable MM:SS or H:MM:SS
viewedCount	All-time view count
publishedAt	ISO 8601 timestamp the talk was published on TED.com
recordedOn	ISO date the talk was recorded
language	ISO 639-1 language code
featured	true for featured talks
curatorApproved	true for curator-approved talks
hasTranslations	true if subtitle translations exist
topics	Array of topic names (e.g. creativity, business, psychology)
thumbnailUrl	Widescreen thumbnail URL
relatedTalkSlugs	Array of related talk slugs
scrapedAt	ISO 8601 UTC timestamp

Empty fields are dropped from every record at every depth.

Input

Parameter	Type	Default	Description
mode	Enum	byUrls	byUrls / byTopic / bySearch / bySpeaker / byPlaylist / browse
talkUrls	Array	["sir_ken_robinson_do_schools_kill_creativity"]	Talk URLs or slugs (mode=byUrls)
topic	Enum	creativity	Curated TED topic slug from a 50-item dropdown (mode=byTopic)
searchQuery	String	—	Free-text search query (mode=bySearch)
speaker	String	—	TED speaker slug or URL (mode=bySpeaker)
playlist	String	—	TED playlist numeric ID or full URL (mode=byPlaylist)
sort	Enum	popular	popular / newest / oldest for mode=browse
minViews	Integer	—	Drop talks with fewer views
minDurationSeconds / maxDurationSeconds	Integer	—	Filter by talk length
language	Enum	(no filter)	ISO 639-1 dropdown of TED's top-30 languages
maxItems	Integer	25	Hard cap (1-1000)

Example input — single talk

{
  "mode": "byUrls",
  "talkUrls": ["sir_ken_robinson_do_schools_kill_creativity"]
}

Example input — multiple talks

{
  "mode": "byUrls",
  "talkUrls": [
    "https://www.ted.com/talks/brene_brown_the_power_of_vulnerability",
    "https://www.ted.com/talks/simon_sinek_how_great_leaders_inspire_action",
    "do_schools_kill_creativity"
  ]
}

Example input — browse a topic

{
  "mode": "byTopic",
  "topic": "creativity",
  "minViews": 1000000,
  "maxItems": 50
}

Example input — short English talks

{
  "mode": "byTopic",
  "topic": "psychology",
  "language": "en",
  "maxDurationSeconds": 600,
  "maxItems": 25
}

Example input — search

{
  "mode": "bySearch",
  "searchQuery": "quantum computing",
  "maxItems": 30
}

Example input — by speaker

{
  "mode": "bySpeaker",
  "speaker": "sir_ken_robinson"
}

Example input — by playlist (Most Popular Talks)

{
  "mode": "byPlaylist",
  "playlist": "171",
  "maxItems": 25
}

Example input — browse most popular talks

{
  "mode": "browse",
  "sort": "popular",
  "maxItems": 25
}

Example output

{
  "recordType": "talk",
  "id": "66",
  "slug": "sir_ken_robinson_do_schools_kill_creativity",
  "url": "https://www.ted.com/talks/sir_ken_robinson_do_schools_kill_creativity",
  "title": "Do schools kill creativity?",
  "speaker": "Sir Ken Robinson",
  "description": "Sir Ken Robinson makes an entertaining and profoundly moving case for creating an education system that nurtures (rather than undermines) creativity.",
  "durationSeconds": 1151,
  "durationFormatted": "19:11",
  "viewedCount": 80129749,
  "publishedAt": "2006-06-27T00:11:00Z",
  "recordedOn": "2006-02-25",
  "language": "en",
  "curatorApproved": true,
  "hasTranslations": true,
  "topics": ["education", "creativity", "psychology"],
  "thumbnailUrl": "https://pi.tedcdn.com/r/talkstar-photos.s3.amazonaws.com/uploads/...",
  "scrapedAt": "2026-05-06T10:42:18Z"
}

Use cases

• Educational platform content discovery — Build curated talk libraries by topic.
• Corporate training catalogs — Index TED talks by length, speaker, or topic for L&D programs.
• Content recommendation engines — Match TED talks to user interests via topics + view counts.
• Speaker / influencer research — Track TED appearances of public figures.
• Academic research — Snapshot communication / public-speaking dataset.

FAQ

Do I need a TED account or API key? No. TED.com pages embed full talk metadata in their __NEXT_DATA__ JSON blob; the actor reads that directly.

How do I find a topic slug? Browse https://www.ted.com/topics — every topic page URL ends in the slug (e.g. /topics/creativity → creativity). Common slugs: creativity, business, education, psychology, technology, science, health, culture, art.

Are transcripts included? Not in this version. Talk transcripts live on a separate /transcript URL and are translated into many languages; capturing them requires an additional fetch per talk and per language.

Why does mode=byTopic make N+1 requests? TED's topic listing returns lighter data per talk. To get full metadata (views, duration, language, etc.), the actor visits each talk's individual URL. Use mode=byUrls if you already have the slugs and want fewer round-trips.

How current is the data? Live — every run hits TED.com at request time. Schedule the actor for daily / weekly refreshes to track view-count growth.

Do I need a proxy? No. TED.com accepts datacenter IPs without restriction.

Limitations

• TED's topic pages return ~16-20 talks per topic; large catalogs need multiple topic queries.
• Transcripts are not yet captured.
• Per-talk comments / reactions are not exposed in the public data.
• Some old talks have sparse metadata (no language, no recordedOn, etc.).