Get Weeve

Transcription in 25 European languages

Weeve uses NVIDIA Parakeet TDT 0.6B v3 to detect and transcribe 25 European languages automatically. There is no language picker. The result includes punctuation, capitalization, and word timestamps.

Drag audio files here

or

Compatibility at a glance

Language detection and transcription are automatic. Audio quality, accents, specialist vocabulary, code-switching, overlapping speech, and short clips can affect the result.

  • Bulgarian
  • Croatian
  • Czech
  • Danish
  • Dutch
  • English
  • Estonian
  • Finnish
  • French
  • German
  • Greek
  • Hungarian
  • Italian
  • Latvian
  • Lithuanian
  • Maltese
  • Polish
  • Portuguese
  • Romanian
  • Russian
  • Slovak
  • Slovenian
  • Spanish
  • Swedish
  • Ukrainian

How it works

Four clear steps from local file to reviewed text.

  1. 01

    Choose the recording

    Select a supported file with speech in a listed language.

  2. 02

    Decode locally

    The browser prepares the audio without sending the recording to a transcription service.

  3. 03

    Detect and transcribe

    The multilingual model chooses the language automatically and returns timed words.

  4. 04

    Review the output

    Check names, specialist terms, punctuation, and speaker labels before exporting.

Product details

Practical details for language support.

No language selection step

The model identifies the spoken language from the recording. You do not need to choose a language before starting, and there is no manual language override.

A clear sample with enough connected speech helps automatic detection. Very short clips or recordings dominated by music and noise provide less evidence.

Multilingual and mixed-language audio

Parakeet is multilingual, but a conversation that repeatedly switches languages is harder than a recording dominated by one language. Review transitions and names carefully.

Speaker separation is a separate acoustic process. It groups voices rather than languages, so speaking a different language does not create a reliable new speaker label.

What to review

Automatic punctuation, capitalization, and timestamps make the first draft easier to use, but they are still model output. Check proper nouns, acronyms, numbers, and domain-specific terms.

Exports reflect the reviewed transcript and current speaker names. They are available as Markdown, plain text, and PDF.

Your recording stays here

Recording bytes, decoded audio, transcript text, and speaker names stay on this page. Speech models download from Weeve’s CDN.

A transcript you can review

Timestamps and speaker labels stay editable before export.

Maya

Let’s confirm the launch checklist before we finish.

Speaker 02

The help article is ready. I will send the final link this afternoon.

Maya

Great. I renamed your speaker label, so the exported notes will use your name.

CopyMarkdownTextPDF
Transcribe now

Frequently asked questions

Practical answers about files, processing, privacy, and results.

Related transcription guides

Continue with product details that match your file.