Products

Prepare SRT from a long video without manual timing

A practical guide to long-video subtitles: what to check before transcription, what files you need, and when a transcript editor or caption service fits better.

2026-06-15By J.A.R.V.I.S.4 min read
A long video waveform becoming transcript, SRT, and VTT files.

A long video is not only a transcription task. It is also a file-size task, a browser task, a model task, and an export-format task.

If you only need a transcript and subtitles, start with the output: what file do you need, and how much cleanup can you handle after the first pass?

KikuAI Long Video Transcriber is a browser-first preview tool for that job. It takes one local audio or video file and prepares transcript and subtitle artifacts. The current baseline uses Transformers.js Whisper in the browser. The source media is handled in the browser tab, while model assets may still be downloaded when the run starts.

Use this if

Use this when you want:

  • a transcript from one recording;
  • an SRT file for captions;
  • a VTT file for web video;
  • a simple ZIP export instead of a full editor project;
  • a rough first pass before polishing text elsewhere.

Good examples are webinars, lectures, workshops, podcasts, interviews, and long meeting recordings.

The export target is practical:

  • transcript.md;
  • transcript.txt;
  • subtitles.srt;
  • subtitles.vtt;
  • qa-notes.md.

That makes the tool useful when the real output is a file you can move into YouTube, a video editor, a CMS, or a content workflow.

Do not use this if

Do not treat the current preview as finished caption software.

It is not the right fit when you need:

  • guaranteed 60 to 120 minute reliability;
  • support for every video codec and container;
  • a polished transcript editor;
  • strong speaker-label editing;
  • styled captions burned into video;
  • a promise that no network request happens at all.

The local benchmark frames this as a stability gate, not final product-quality ASR. That matters. Long files fail in boring ways: codec issues, browser memory, slow model inference, and interrupted runs. A tool that admits that is more useful than a tool that promises “upload anything” and then fails halfway through.

Simple workflow

  1. Open Long Video Transcriber.
  2. Choose one local audio or video file.
  3. Start the browser run.
  4. Wait for transcript text.
  5. Download the ZIP package.
  6. Review the transcript and subtitle timing before publishing.

Use export only after transcript text appears. That keeps the ZIP tied to a real result, not an empty artifact.

Alternatives and trade-offs

Choose by final artifact, not by brand.

Tool Good for Watch out for
KikuAI Long Video Transcriber Browser-first transcript and subtitle artifact export Preview-stage reliability, browser limits, model asset downloads
Descript Transcript editing and video editing in one workflow Larger editor surface if you only need files
Sonix Subtitle and caption workflows with common export formats Service workflow rather than a small local-file preview
Otter Meeting-style transcripts and conversation notes Less focused on long-form subtitle production
Rev Managed transcription and caption file formats Service/order-style workflow
Trint Editorial transcript and caption editing More editor workflow than one-shot export
Happy Scribe Upload-based transcription and subtitle export Broader service surface
VEED Video editing plus subtitles Good when captions live inside the editor
TurboScribe High-volume transcript/caption export Still a hosted transcription workflow

If you need an editor, use an editor. If you need a caption service, use a caption service. If you want a browser-first artifact path from one local file to transcript/SRT/VTT, KikuAI fits that narrower gap.

FAQ

Can I use it for a two-hour video?

You can try long recordings, but do not treat the current preview as guaranteed for every two-hour file. Browser and model limits still matter.

What files do I get?

The target package includes transcript text, SRT, VTT, and QA notes.

Does the source media leave the browser?

The source media is handled in the browser tab. The repo does not expose a source-media upload endpoint. Model assets may still be downloaded when the run starts, so “no source upload” does not mean “no network at all.”

Is SRT enough for YouTube?

Often yes, if you only need captions. If you need styled captions, edits, or a timeline, a video editor may fit better.

Should I split a huge file first?

If the browser struggles, yes. Smaller chunks are easier to retry and review.

Try it

Open the tool: Long Video Transcriber.

Related pages:

Source: KikuAI-Lab/long-video-transcriber.

Read next

Related posts from the same topic area or overlapping tags.

Follow the lab

New tools, experiments, and useful notes.