Why this page preselects Ukrainian
Ukrainian and Russian share an alphabet and a lot of vocabulary. When a program has to guess the language from the first seconds of audio, a Ukrainian recording can be taken for Russian, especially if it is short, starts with a name or a greeting, or if the speaker uses Russian loanwords. The result is then written in Russian spelling or with Russian words, which is not what you want.
This page sets the language to Ukrainian in the upload box before you start. That keeps the output in Ukrainian. If you have a recording that is mainly Russian, change the setting to Russian in the same box.
Who needs Ukrainian transcription
Since 2022 many Ukrainians live in Poland, Germany, Czechia, the UK, Canada and the US, and a lot of recordings move across borders. Typical uploads include interviews by journalists and NGOs, oral history and documentation projects, video messages from family, lectures from Ukrainian universities that teach online, church services, and voice notes from colleagues back home. There is also a long-established Ukrainian diaspora in Canada and the US with family tapes and community radio archives.
Ukrainian is one of the languages the speech model is strong in. Clear speech usually comes out as readable text. Field recordings with wind, traffic or several people talking at once need more editing. If you work on sensitive material, note that files are never used to train AI models and you can delete them at any time.
Getting good Ukrainian audio to text from field recordings
Many Ukrainian recordings are made on a phone in less than ideal places: a kitchen with the television on, a car, a shelter, a street. A few habits help the transcript more than anything you can do afterwards. Hold the phone or place it close to the person speaking, ask people not to talk over each other, and turn off music or the TV if you can.
Upload the original file. Voice messages from Telegram, Viber or WhatsApp and videos straight from the phone can be uploaded as they are, and the audio is extracted on our side. Converting them first only loses detail. Long interviews are fine too, because uploads are sent in parts and Retry continues where a dropped connection stopped.
Ukrainian letters, mixed speech and names
The transcript uses the Ukrainian alphabet, including і, ї, є and ґ, and the apostrophe in words like п'ять and сім'я. Look at these points when you review it.
- Mixed Ukrainian and Russian speech (often called surzhyk) is common in everyday conversation. With Ukrainian selected, Russian words may be written in Ukrainian spelling or replaced by a Ukrainian word. If you need the exact words for a quote or for evidence, check those passages against the audio.
- Conversations that switch fully between the two languages are harder than a recording in one language. Pick the language spoken most.
- Names of people, villages and military or administrative units: check spelling against your sources.
- English words such as app names may appear in Latin letters or in Cyrillic.
- Numbers and dates can appear as digits or words.
Exports and English versions
TXT, DOCX (Word), SRT, VTT and CSV keep Ukrainian characters exactly. The direct PDF export supports Western European characters only, so for a Ukrainian transcript the PDF option opens a print-ready page that you save as PDF from your browser.
RocketScribe gives you the Ukrainian text, not an English translation. Use the Google Translate link on the transcript when a colleague needs to read it in English. For an English summary, use the AI prompt export, which copies the transcript with a prompt for ChatGPT or Claude. Machine translation is fine for understanding content, but for court, asylum or medical use you should have a qualified translator check it.
Speaker labels for interviews with several people are available on paid plans or with a Minute Pack. SRT and VTT export the Ukrainian text as subtitles for a video.