Formerly NovaScribe — same team, same product, refreshed name. Read the announcement →

Transcribe and Translate Audio

One upload, two outputs: an accurate transcript in the original language plus a translated version in 133 languages. AI-powered, no separate translation tools needed.

VexaScribe (formerly NovaScribe) is the AI transcription tool that does both: transcribe audio in 99 languages using OpenAI's Whisper Large-v3, then translate the resulting transcript into any of 133 languages using Google Translate — all in one workflow, included on every paid plan at no extra cost. Upload a podcast in Spanish; export an English transcript and English subtitles. Record a German interview; get French, Japanese, and Portuguese versions for international publication. Most transcription tools (Otter, Rev, Descript) are transcription-only — translation requires a separate tool, separate account, separate workflow, and (usually) a separate bill. VexaScribe combines them: Whisper handles speech-to-text in the original language, Google Translate handles text-to-text into your target language. Free tier includes 30 minutes of transcription; paid plans start at $2/month with translation included on all of them.

99 transcription languages133 translation languagesTranslation included freeOne workflow

Two AI Models, One Workflow

Most tools stop after transcription. We continue through translation — automatically.

Audio file
MP3, WAV, MP4, MOV, etc.
OpenAI Whisper Large-v3
Transcription · 99 languages
Original-language transcript
Speaker labels, timestamps
Google Translate
Translation · 133 languages
Translated transcript + subtitles
SRT, VTT, DOCX, TXT, JSON

You never copy-paste the transcript into Google Translate, never sign up for a translation API, never manage two accounts. The whole workflow happens automatically when you click “Translate to [language]” in the editor.

Common Workflows

Real use cases where transcription + translation in one tool saves hours per week.

Spanish podcast → English subtitles

Reach English-speaking audiences without recording twice. One source episode, two language versions.

German interview → 4 European languages

Transcribe a German interview once, publish French, Italian, Portuguese, and Dutch versions.

Japanese course → English materials

Open Japanese course content to global learners. Transcribed lecture notes, translated study guides.

Multilingual sales calls → English summary

Customers speak their language; sales reps read transcripts in English. No interpreter needed.

Documentary footage → unified script

Editors working across language barriers. Transcribe each language, translate to a common language for editing.

Conference talk → 8 languages

One recorded talk, eight translated transcripts. Make a single conference session globally accessible.

99 Languages for Transcription, 133 for Translation

Audio is transcribed in 99 languages with automatic detection. Transcripts can be translated into 133 target languages.

99

Transcription languages

Tier 1~5% Word Error Rate
EnglishSpanishFrenchGermanItalianPortugueseDutchPolishRussianJapanese
Tier 2~8–12% WER
ArabicChineseKoreanHindiTurkishVietnameseThaiIndonesianHebrewCzechSwedishNorwegianDanishFinnishGreekUkrainian

+ 73 more including Welsh, Swahili, Filipino, Bengali, Punjabi

133

Translation languages

All 99 transcription languages plus 34 additional translation-only languages:

AlbanianAmharicArmenianAymaraBasqueBhojpuriBosnianBulgarianCebuanoChichewaCorsicanEsperantoFrisianGalicianGeorgianGuaraniGujaratiHaitian CreoleHausaHawaiianHmongIgboKannadaKhmerKinyarwandaKurdishKyrgyzLaoLatinLatvianLingalaLithuanianLugandaLuxembourgishMacedonian

Powered by Google Translate. Translation quality is highest for major Western European, East Asian, and Arabic pairs.

How Accurate Is the Translation?

Translation quality depends on the language pair and the content type. Major Western European languages (Spanish, French, German, Italian, Portuguese) translate to and from English at very high quality — Google Translate has been training on these for over 15 years.

Less common pairs (Khmer ↔ Welsh, for example) have noticeably lower fluency. Technical content with industry jargon may need human review before publication.

VexaScribe's role is the workflow: we hand the transcribed text to Google Translate and present the result. The translation quality is whatever Google Translate produces — we don't claim to improve it. Our value is eliminating the copy-paste step and the second-tool subscription.

Recommendation: For publication-grade content, treat the translated transcript as a strong first draft. Review it manually — especially for technical terms, idioms, and culturally-specific phrasing. The side-by-side editor makes manual refinement fast.

What You Get

Two complete transcript versions plus everything else VexaScribe normally produces.

Original-language transcript

Full transcript in the language the audio was recorded in. Export to TXT, DOCX, SRT, VTT, or JSON.

Translated transcript

Same export formats, in your target language. Pick one or more — generate English, French, Spanish from a single source.

Side-by-side editor

View original and translation in two columns. Edit either version. Useful for verifying translation accuracy or polishing wording.

Translated subtitles

Word-level timestamps preserved through translation. Upload directly to YouTube as a translated caption track. SRT and VTT supported.

Why Not Just Use Google Translate Separately?

You can. The math is whether the manual workflow is worth it for your use case.

AspectWith VexaScribeWithout VexaScribe
WorkflowOne upload, both versions in one placeTranscribe in tool A, copy-paste to Google Translate, paste back into subtitle tool
Word-level timestampsPreserved through translation — translated SRT exports correctlyLost during copy-paste; rebuild SRT timestamps manually
PricingTranslation included free on all paid plans (from $2/mo)Google Translate API requires separate account + usage-based billing
Speaker labelsCarry over to translated version automaticallyLost during copy-paste; manually re-add for each language
Effort per useAutomatic — set it and forget itManual workflow every time, every language pair

Rule of thumb: if you transcribe + translate once a year, the manual workflow is fine. If you do it weekly (podcast publisher, newsroom, course creator), VexaScribe saves hours per use.

Translation Included on Every Paid Plan

No upcharge, no separate translation account, no usage-based fees. Just unlimited translation across 133 languages.

Starter

$2/month

200 min/month

Translation included

Basic

$5/month

1,000 min/month

Translation included

Pro

$10/month

2,500 min/month

Translation included

Frequently Asked Questions

How does VexaScribe transcribe and translate audio in one step?

VexaScribe (formerly NovaScribe) combines two AI models in one workflow. Step 1: OpenAI's Whisper Large-v3 transcribes your audio into text in the original spoken language (99 languages supported, with automatic detection). Step 2: Google Translate translates the resulting transcript into your target language (133 languages available). Both happen automatically — you upload once, get both versions in your dashboard within 5-10 minutes for a one-hour file.

How many languages does VexaScribe support for translation?

VexaScribe translates transcripts into 133 languages using Google Translate. This includes all 99 languages we transcribe plus 34 additional translation-only languages — Welsh, Bengali, Punjabi, Tamil, Urdu, Albanian, Armenian, Esperanto, Hawaiian, Khmer, Lao, Latin, Macedonian, and many more.

Is translation included free with the transcription plan?

Yes. Translation is included on every paid plan at no extra cost — Starter ($2/month), Basic ($5/month), Pro ($10/month), and Studio ($20/month) all include unlimited translation. The 30-minute free trial also includes translation. There's no separate translation account, no API key, no usage-based fees.

Can I translate to and from any language pair?

Yes. You can translate from any of the 99 transcription languages to any of the 133 translation languages. Common pairs (Spanish to English, English to French, German to Italian) work very well because Google Translate has been trained on these for over 15 years. Less common pairs (Khmer to Welsh, for example) have lower fluency.

What's the translation quality like for less common languages?

Translation quality depends on the language pair and the content type. Major Western European languages (Spanish, French, German, Italian, Portuguese) translate to and from English at very high quality. Asian languages (Japanese, Korean, Chinese) and Arabic perform well. Less common pairs and technical content with industry jargon may need human review. VexaScribe doesn't claim to improve on Google Translate's output — our value is eliminating the copy-paste workflow and the second-tool subscription.

Can I export translated subtitles (SRT/VTT)?

Yes. Translated transcripts can be exported as SRT or VTT subtitle files with the original word-level timestamps preserved. Upload a Spanish video, transcribe to a Spanish SRT, then export an English translated SRT for the same video — both files have correctly synced timestamps. Perfect for international YouTube uploads.

Does the speaker detection carry over to the translated version?

Yes. Speaker labels (Speaker 1, Speaker 2, or your custom names) are preserved through translation. If you rename speakers in the original transcript, those names appear in the translated version too. Useful for multilingual interviews where you need consistent speaker attribution across language versions.

What's the difference between transcription and translation?

Transcription converts spoken audio into text in the same language — a Spanish recording becomes Spanish text. Translation converts text from one language to another — Spanish text becomes English text. VexaScribe does both: transcribe Spanish audio to Spanish text, then translate that Spanish text to English (or any of 132 other languages).

Why not just use Google Translate separately?

You can. The downside: you have to transcribe in tool A, copy-paste the transcript into Google Translate, paste back into your subtitle tool, and rebuild timestamps manually. With VexaScribe, it's one upload and the entire workflow happens automatically — speaker labels and timestamps preserved through translation, exportable as SRT directly. If you do this once a year, the manual workflow is fine. If you do it weekly, VexaScribe saves hours.

Can I edit the translation manually?

Yes. The built-in editor shows the original-language transcript and the translated transcript side-by-side. You can edit either version, add corrections, fix machine-translation errors, or refine wording. Both versions update independently so you can preserve the accurate original while polishing the translation.

Reach Global Audiences with One Upload

30 minutes free, no credit card required. Start with one file and see both languages in your dashboard within minutes.