YouTube Transcript Downloader – TXT, SRT & JSON

Getting a transcript from YouTube takes less than ten seconds and works for standard videos, YouTube Shorts, past live replays, premieres, and unlisted videos — on desktop and mobile.

1. Copy the YouTube video URL. Open the video and copy the link from your browser's address bar, or tap Share → Copy link in the YouTube app. Every URL format works: youtube.com/watch?v=..., youtu.be/..., youtube.com/shorts/..., and youtube.com/live/....
2. Paste the URL into the input box above and click Get Transcript. The tool asks YouTube's caption API for every available track (manual captions, auto-generated, translated) and parses the results with timestamps.
3. Download or copy the transcript. Pick TXT for notes and reading, SRT for subtitle files that work in every video player and editor, or JSON for scripts and AI pipelines. If multiple languages exist, use the language dropdown first.

Unlike tools that try to parse the transcript from the on-screen Show Transcript panel (which truncates long videos and strips timing details), this tool downloads the raw caption track directly from YouTube's servers. That means you get every cue, accurate millisecond timestamps, and no truncation — even for multi-hour documentaries or lecture recordings.

Tips for the best results

• When multiple tracks are available, manual captions are always higher quality than auto-generated speech recognition.
• Need the transcript for a blog post or video script? Download TXT, paste into your editor, and remove the timestamps in one search-and-replace.
• Editing the video and want to reuse captions? SRT imports natively into Premiere Pro, CapCut, DaVinci Resolve, Final Cut, iMovie, and every editor worth using.
• Working with AI tools or LLMs? JSON gives you start_ms, end_ms, and per-line text for clean programmatic access.

A YouTube transcript is the written version of everything spoken in a video, matched to timestamps showing exactly when each line was said. YouTube gets transcripts from two sources:

• Manual captions — uploaded by the creator or a translator. These are always the gold standard: accurate, properly punctuated, and corrected for jargon and names.
• Auto-generated captions — produced by YouTube's ASR (automatic speech recognition) system. Quality varies; English, Spanish, and major languages on high-audio videos can be surprisingly good, while niche languages, strong accents, heavy background music, and technical vocabulary often produce unusable output.

Creators, students, podcasters, researchers, and editors download transcripts for a wide range of legitimate uses: quoting from videos with citations, turning conference talks into blog posts, making notes from lectures, pulling quotes for documentaries, extracting talking points for summaries, feeding speech into AI for analysis or translation, accessibility for deaf or hard-of-hearing viewers, and recovering lost scripts when the original project files are gone.

If the tool returns No transcript available, it means YouTube has no publicly accessible caption track for that video. The most common reasons, in order:

• The creator disabled captions entirely. A small number of channels turn off all captions in their YouTube Studio settings. No tool can work around this.
• Private, members-only, or heavily age-restricted videos. These never expose captions on the public API.
• Very short or very new uploads. Under 30 seconds YouTube often skips auto-captions entirely. For a fresh upload, wait 5–30 minutes after publishing before retrying — ASR processing runs asynchronously.
• No recognisable speech. Music-only videos, silent clips, and videos with only non-speech audio get skipped by the auto-caption system.

Unlisted and public videos always attempt caption generation. If you see one language missing but another exists, try the language selector — sometimes YouTube only generates captions in the video's original spoken language.

One of the most popular uses for a transcript downloader is feeding video content into large language models like ChatGPT, Claude, Gemini, or custom RAG pipelines. The JSON download is designed for this workflow because it:

• Keeps timestamps. Every cue has millisecond precision so your summarizer or quote bot can reference exact moments in the video.
• Preserves structure. Line breaks follow YouTube's caption segmentation, which roughly aligns with speaker turns and sentence boundaries.
• Encodes cleanly. Unicode is escaped properly and the output is valid JSON, parseable by every language without text pre-processing.

A typical AI workflow: download JSON, concatenate the text fields into a single prompt (optionally prepending timestamps), ask the model to produce a summary, key takeaways, chapter list, or Q&A set, then append start_seconds citations to each answer so readers can jump straight to the relevant part of the video.

Transcripts are derivative works of a video's audio, and copyright normally belongs to the video creator or publisher. What you can do with a downloaded transcript is roughly the same as what you can do with the video itself:

• Generally fine: private study, personal notes, accessibility, research, quotation with attribution, internal meetings, and your own videos if you downloaded your own transcript.
• Often defensible as fair use (jurisdiction-dependent): commentary, criticism, reviews, news reporting, and educational materials that quote the transcript to discuss the original video.
• Risky or clearly infringing: republishing an entire transcript as your own, uploading it to another platform, turning it into a spam blog post, or using it commercially without permission.

For creators downloading their own channel's transcripts — which is one of the most common use cases when project files were lost — you, of course, retain full rights. This tool is a convenient way to recover work YouTube already processed for you.

This section is general information, not legal advice. Copyright rules vary by country.