Batch YouTube Transcripts on Mac

Why batch transcripts are usually painful

Most YouTube transcript tools were designed for one video at a time. Browser extensions live on a single tab. Web tools paste a URL, give you a result, and reset. The few tools that claim batch usually mean "paste a few URLs and pray." Three things go wrong on a real batch:

• YouTube rate-limits the requests, and the tool gives up instead of backing off and retrying.
• The tool runs in a browser tab, and the tab needs to stay focused, so the laptop cannot sleep and you cannot switch apps.
• Failures are silent. You end up with a partial result and no clear list of what failed and why.

Vidgest's batch system was rebuilt to address all three.

How Vidgest's batch queue works

1. Paste, do not pick

Open Vidgest, paste a newline-separated list of YouTube URLs, and press start. There is no per-URL setup. The queue accepts plain video URLs and shortened youtu.be URLs in the same paste.

2. Adaptive rate-limit handling

YouTube throttles repeated requests. Vidgest's fetcher uses a circuit-breaker pattern. When a throttling signal is detected, the queue pauses, waits with backoff, then resumes. The breaker has graduated recovery (it needs multiple successes before fully closing), which prevents the queue from hammering YouTube right after a throttle clears. You see live status on each URL.

3. Background execution

Vidgest is a native macOS app, not a browser tab. The batch keeps running when you switch to Slack, lock the screen, or work in another app. You get a notification when the queue completes.

4. Clear failure reporting

Each URL in the queue ends with a specific status: success, video not found, captions disabled, or rate-limited. Successes are stored locally; you do not lose them if a later item fails. You can retry the failed subset without re-fetching the successes.

How big can a batch be?

No hard cap. The app has been stress-tested with batches of 750 URLs in a single session and stayed stable. A typical 50 to 100 URL batch finishes in tens of minutes depending on YouTube's current rate-limit posture and the length of each video. The pacing copy in the app changes as time passes ("Fetching next..." early on, "Working through your list..." after 2 minutes, "Still going, feel free to switch apps" after 5 minutes) so you know it is still alive even on a multi-hour run.

What this is useful for

• Research catch-up: A folder of saved videos becomes a folder of searchable transcripts.
• Course playlists: Pull every lecture from a YouTube playlist into Markdown for your notes app.
• Competitive analysis: Bulk-fetch a competitor's channel so you can search what they have actually said.
• Podcast index: Subscribe to a podcast that publishes on YouTube and bulk-grab the back catalog.
• Subtitle prep: Pro export to SRT/VTT for use as subtitle files on your own edits.

After the batch: AI analysis

Once the queue finishes, every transcript sits in your local library. AI analysis is one-at-a-time today. The free tier gives 3 AI analyses per day. Pro gives 500 per month. Max ($49.99 one-time) lets you bring your own Google Gemini, Anthropic Claude, or OpenAI API key, in which case there is no Vidgest-side cap.

Privacy on a batch this big

Vidgest stores every transcript in a local SQLite library on your Mac. No cloud sync, no account, no server-side history. AI analysis is the only step that leaves the machine, and only when you explicitly run it. A 500-video batch is a 500-row local table. Nothing is uploaded by default.

Frequently asked questions