YouTube Speech & Presentation Transcript

What “publish accurate excerpts with times” actually requires

Speeches are quoted widely. Timed ASR helps, but verify every excerpt against the audio.

Only authorized recordings. Political and corporate talks often have strict rules.

Crowd noise hurts accuracy.

Unlike a generic file upload, this path starts from a YouTube URL aimed at comms teams and students. That is useful when the only artifact you hold is the watch link, but it also means refusals for playlists, private uploads, age gates and live events still apply.

Expect to edit: Whisper drafts the speech so you can publish accurate excerpts with times, yet names, numbers and jargon still need a human pass. If you already have an MP4 on disk, video to text may be simpler than re-hosting on YouTube first.

Speeches and keynotes are usually one continuous voice with rhetorical structure. Use plain-text mode after you mark applause/aside stretches — distinct from interview Q&A labeling.

How to run the speeches workflow

1. Open YouTube Speech & Presentation Transcript intent: drop the public YouTube URL that matches this speeches use case into the tool above.
2. Decide whether speeches needs a fixed spoken language or auto-detect for this recording.
3. Tick rights specifically for this speeches source — ownership or permission, not “it was public.”
4. Confirm rights, upload the soundtrack (yt-dlp on your computer or the one-shot helper), then start Whisper; credits follow ordinary transcript pricing.
5. When text appears, tune timestamps for publish accurate excerpts with times, then copy only the spans this workflow needs.
6. Continue in your editor, notes app or LMS; return to YouTube transcript if you need another permitted link.

Example

Practical tips for speeches

• Prefer official channels.
• Verify numbers twice.
• For speeches, favour recordings where the main voice sits above music — ASR quality tracks the mix.
• If speeches work fails on a private or live URL, wait for a public finished upload or use a file on video to text.
• Keep a local note of the video id beside every speeches paste for provenance.

Boundaries for speeches

• YouTube Speech & Presentation Transcript still accepts only one finished public video URL — never a playlist shelf or channel home.
• Clips over two hours are out of range for speeches; split in Studio or export parts to video to text.
• The speeches output is speech recognition text, not a dump of YouTube’s existing caption track.
• Instagram, TikTok and Vimeo links are unsupported here; speeches is YouTube-only.
• mydubly will not give you the media file after a speeches run — only the words.

Using the tool on this page

• Paste one public YouTube video link (not a playlist or channel). Confirm you have the right to transcribe it before you start.
• Whisper recognises speech in about 100 languages, or detects the language automatically. Videos can be up to 2 hours.
• You get text you can show with timestamps or as plain copy. The soundtrack is used only to make the transcript, then deleted within 30 minutes — mydubly does not hand you the video or audio file.
• Pricing matches other transcripts: 1 credit per minute (0.1¢), minimum 5 credits per file. More detail is on the YouTube transcript generator page.