How good is a Bosnian draft
The speech model has seen far less Bosnian than English or German, so plan on a full review. A clear recording often gives a readable draft with most common words right, but names, places, dialect words and specialist terms need checking every time.
One speaker near the microphone in a quiet room works best: a lecture, a dictated text, an interview with a clip-on mic. Kitchen-table conversations, calls on speakerphone, several people talking over each other, or speech over music give weaker output. Even then, the timestamps help you find your place while you correct the text.
A practical way to work: play the file once while following the highlighted line, fix obvious errors on the spot, then do a second pass only for names, numbers and quotes. For text you plan to publish, cite in research or hand to a court or an authority, have a native speaker read the final version. Small spelling differences matter in Bosnian, because they show which standard a text follows.
Bosnian, Croatian, Serbian and auto-detect
Bosnian, Croatian and Serbian are mutually intelligible, and the model cannot always tell them apart. On auto-detect, Bosnian speech is often labeled as Croatian or Serbian. A Serbian label can mean the whole transcript comes out in Cyrillic, and a Croatian label can pull word choices toward Croatian forms. Short voice notes and clips that start in English are the most likely to be misjudged.
This page preselects Bosnian to reduce that. If you get Cyrillic you did not want, or the text reads like another standard, check the language setting and run the file again with Bosnian selected. Even with Bosnian selected, some word forms may follow Croatian or Serbian usage, so look out for them in review.
Bosnian spelling: what to check
Bosnian is usually written in Latin letters, and the transcript follows that. These are the details a review most often has to fix.
- Diacritics: č and ć, dž and đ are easy to confuse because they sound close in fast speech, and ž, š or đ may lose their marks.
- Ijekavian forms such as mlijeko, djeca and vrijeme can come out as Ekavian (mleko, deca, vreme).
- The h that Bosnian keeps in words like kahva, lahko, sahat or hudovica may be dropped.
- Turkish, Arabic and Persian loanwords common in Bosnian speech (avlija, merhamet, sevdah, insan) may be misheard or replaced with a similar common word.
- Personal and place names, especially from smaller towns, need checking against a reliable source.
The diaspora and mixed speech
Bosnians abroad, in Germany, Austria, Sweden, Slovenia, Switzerland, the US and Australia, often mix German, Swedish or English words into Bosnian, and their children may switch languages mid-sentence. Choose the language by what dominates the file. A call that is mostly German with a few Bosnian sentences works better with German selected.
Common uploads include family recordings and wedding videos, grandparents describing life before and during the war, oral history and research interviews, journalist interviews, university lectures, mosque and community talks, and YouTube videos that need Bosnian subtitles.
Uploading and exporting
WhatsApp and Viber voice notes, phone recordings and MP4 or MOV videos upload directly; audio is pulled out of video automatically. The transcript comes in timestamped paragraphs, and the player highlights the line being spoken. Click a line to replay it and correct the text in place.
Export TXT, DOCX, PDF, SRT, VTT or CSV, with timestamps and speaker names on or off. For an English reading, open the Google Translate link on the transcript, or use the AI prompt export with ChatGPT, Claude or Gemini. Speaker labels are available on paid plans or with a Minute Pack. Your files are never used to train AI models.