- file
- youtube-transcription.transcript
- speakers
- You, vocateca
- address
- vocateca.com/en/youtube-transcription
- duration
- 00:04:16 · 588 words
Read any YouTube video — timestamped, on your Mac
Opening
Paste a link. Player on the left, timestamped transcript on the right, kept in sync. Captions load instantly; nothing is uploaded.
Paste, read, click to seek
Instant captions. Drop a URL and read a scrollable, timestamped transcript in seconds — before any audio is processed. Click any line. The player jumps to that moment; a floating play button follows your scroll so you always start “from here”. Speaker labels on demand. One button runs the on-device pipeline — Parakeet-TDT plus diarization — for a speaker-labelled transcript.
Not just another transcript site
Most online tools post a URL to a server and hand back plain text. Vocateca is a native macOS app built for working with a lot of audio.
| Vocateca | Typical online tools | |
|---|---|---|
| Where it runs | On your Mac — URL and audio never leave it | Server-side — your link and audio hit their servers |
| Timestamps | Core: every segment timed, click to seek | Often optional or plain-text only |
| Speaker labels | Yes — on-device diarization, carried into CSV | Rare |
| No captions? | Falls back to real on-device transcription | Usually fails — captions only |
| Auto-caption cleanup | Collapses YouTube's rolling-window duplicate lines into clean, correctly-timed segments | Paste raw — every line repeated |
| Export formats | md, txt, srt, vtt, csv, json, okf | Plain text, sometimes SRT |
| Beyond one video | Playlists, channel subscribe, auto-download | One video at a time |
| Automation | Built-in CLI + MCP transcript command | Web UI or a paid API |
| Account | None — no sign-up, no credits | Often sign-up or metered credits |
What we leave out, on purpose. No cloud chat, no server API, no AI summaries — off-model for a native, private app. Vocateca does one job well: clean, structured text out of the audio you follow.
What leaves your Mac for this source
These entries come from the complete list. The entries that belong to no single source, such as updates and speech models, are on the complete list too. The complete list
| # | What | Goes to | Only when |
|---|---|---|---|
| When you add, search or refresh what you follow4 entriesonly when you act | |||
| 01 | YouTube videos and captions | Fetched by yt-dlp, which ships inside the app.youtube.com | You add a video or playlist, or a channel refreshes. |
| 02 | YouTube search and channel feeds | The search results page, and each channel’s public feed.youtube.com/results · youtube.com/feeds/videos.xml | You search YouTube, or a channel you follow refreshes. |
| 03 | YouTube player | The YouTube Explorer plays the video in YouTube’s own player. The page around it comes from a small server inside the app on 127.0.0.1, which only your Mac can reach.youtube.com/iframe_api | You open a video in the YouTube Explorer. |
| 04 | Artwork and thumbnails | Loaded once, then kept in a cache on your Mac.the server that hosts each image | A show, channel or search result is shown. |
| The first time a feature needs it1 entryonce | |||
| 05 | yt-dlp, fallback only | A pinned version. Only used if the copy inside the app can’t run.github.com | Only if the bundled yt-dlp fails. |
More than a single video
Every format, speakers included. Subtitles as VTT/SRT, structured segments as CSV/JSON, LLM-ready Markdown. CSV carries a speaker column when diarized on-device. All export formats
Playlists, channels, daemon. Paste a playlist or a CSV of links to transcribe in bulk — or subscribe to a channel and let a daily daemon do it.
On-device, no account. Captions from YouTube, transcription local on Apple Silicon, transcript stays on your Mac unless you export it. How privacy works
Send a video from YouTube in one click
Install the Vocateca browser extension and a small transcript button appears under every YouTube video. Click it and the video is sent straight to the Vocateca app on your Mac, which extracts the transcript for you — no copy-paste.
One click. A transcript button sits right next to Like and Share on every watch page.
Straight to the app. The link opens in the running Vocateca app — the extension itself sends nothing to any server.
You choose. Open the transcript right away, or silently queue it without leaving your browser.
Scriptable, and usable by AI agents
The same extraction is a command-line tool and — automatically — an MCP tool your AI assistant can call.
Extract a timestamped transcript
$ vocateca transcript "https://youtube.com/watch?v=…" --format vtt
WEBVTT
00:00:00.000 --> 00:00:04.120
Hello and welcome
…Vocateca's MCP server is derived from its CLI, so transcript shows up as a tool in Claude with no extra setup.
End of transcript
Read YouTube instead of watching it