Universal workbench

Matroska speech to Timestamped transcript JSON

Transcribe Matroska speech to Timestamped transcript JSON entirely on this device with a revision-pinned Whisper model.

Device-onlyNo uploads. Ever.

Pipelines, branches, presets, naming, and detailed controls

Display

Text sizeDensity High contrast Monospace editor

Quick start

JSON → YAML WebP → PNG SVG → React JSX Plain text → Base64

From

Help me choose the best output Suggested: MP4

What matters most?

Recommendations use the registered target format and tested fidelity classification. They never inspect or upload your content.

01Input

Drop files or a folder hereChoose Matroska files or a folder — batch input is welcome

Choose files Choose folder

or paste content

02Configure & run

MatroskaTimestamped transcript JSON

structural conversion · single or batch

Conversion preflightSize, memory, engine, execution, and fidelity before you start

Input: 0 B
Estimated peak memory: 256 MiB
On-demand download: Up to 78 MiB on first use
Execution: Responsive local worker
Fidelity: No lossy step declared

What to provideDrop one or more Matroska files. Codec support is checked by the on-demand media engine.What you’ll getProduces full text plus timestamped transcript segments.

Spoken languageAutomatic detection works across supported languages. Choosing a language can improve accuracy and speed.Speech taskTranscribe keeps the spoken language. Translate asks multilingual Whisper to produce English.

Output naming & batch behavior

Filename templateAvailable variables: {output} (converter’s meaningful name), {name}, {target}, {ext}, {index}. Duplicate names receive a safe numeric suffix.

Exampleoutput.transcript.json

Pipeline 1 linear · 0 branches

Why use A → B → C?

A pipeline feeds each result directly into the next compatible converter. Use one when the destination needs an intermediate representation or when you want to inspect and normalize that middle format.

JSON → CSV → XLSX turns records into rows before building a workbook.
Markdown → HTML → DOCX resolves markup before creating a document.
SVG → PNG → PDF intentionally flattens vector artwork before placing it on a page.

Only type-compatible next steps are offered. Lossy steps are marked before you run them.

1Matroska speech to Timestamped transcript JSONstructural

Add content above to start. Nothing leaves this browser.

Processing happens locally. Your file bytes are never sent anywhere.

About this conversion

How to convert Matroska to Timestamped transcript JSON

Before you start

Drop one or more Matroska files. Codec support is checked by the on-demand media engine.

Accepted: .mkv · video/x-matroska

Choose or drag one or more Matroska files onto the drop zone.
Confirm the detected format and adjust the available settings. Mixed-format batches are flagged before conversion.
Select “Convert on this device.” Produces full text plus timestamped transcript segments.

How this converter behaves

Pinned OCR or speech-recognition models run locally after an explicit engine download. Recognized text, timing, and confidence data remain in this browser tab.

What happens to your data?

The conversion engine runs in this browser tab. OmniCastConverter has no upload API and never sends your content, file name, or converted result to a server.

Fidelity and limitations

This route is classified as structural. The converter preserves meaning and structure supported by both formats.

The quantized Whisper Tiny model prioritizes download size and local speed; difficult audio can benefit from a larger offline desktop model.