Learn from TED Talks
How to use LangKey on TED Talks — full talk subtitles with timestamp sync and 7 practice modes.
Learn from TED Talks
LangKey natively supports TED Talks. Open any talk on ted.com and LangKey will auto-extract multi-language subtitles, displaying the full transcript with timestamp sync in the practice panel.
How Subtitles Are Captured
TED uses hls.ted.com to push WebVTT subtitle streams. LangKey intercepts these requests in the background and parses them into unified SubtitleItem[], sharing the same practice pipeline as YouTube and Bilibili.
Supports:
- All TED multi-language tracks
- Human-translated and auto-generated (auto-generated marked in UI)
- Talk metadata (title / speaker / duration / category)
7 Practice Modes
Identical to YouTube:
- Input Mode: Word-by-word input of original sentences
- Option Mode: Shuffle sentence order
- Fill-in Mode: AI-generated cloze at your level
- Listening Mode: Listen and pick
- Shadowing: Read after the speaker, AI scoring
- Dictation Mode: AI generates dictation at CEFR level (new in v2.4.0)
- AI Conversation: Scenario-based, AI plays a role (v2.5 roadmap)
Recommended Workflow
- Open ted.com/talks
- Pick a topic you're interested in (start with 5-10 minute talks)
- Launch the LangKey subtitle panel
- Switch to Shadowing mode, repeat 3-5 sentences
- Wrong words auto-enter your vocabulary, reviewed via SRS
Notes
- TED embeds on other sites are not supported; open ted.com directly
- Live streams / talks without subtitles will show empty state
- Auto-generated subtitles vary in quality; prefer talks with "Transcripts" badge
Feedback
If TED subtitle parsing has issues, please report at GitHub Issues with the talk URL.