Persian and Farsi audio to text: an honest view
Farsi is the name of the language in Persian, so Persian and Farsi mean the same thing here. Select Persian in the language list for both. The speech model has much less Persian training data than it has for languages like Spanish or German. Treat the transcript as a fast first draft that you review, not a finished document.
Clear audio makes the biggest difference. One speaker close to the microphone in a quiet room gives a usable draft that needs moderate editing. A phone call, a family gathering with several voices or a recording with music underneath will need much more correction. Check every name, place and technical term before you rely on the text.
Reviewing is quicker than typing from scratch. Play the file, follow the highlighted line, and click any line to hear it again before you correct it. Voice messages saved from WhatsApp or Telegram can be uploaded directly, and so can phone recordings and videos, with no conversion step.
Iranian Persian, Dari and Tajik
Most recorded Persian is Iranian Persian, and that is what the model handles best. Dari, spoken in Afghanistan, is the same language with different pronunciation and some different vocabulary. Select Persian for Dari recordings as well, and expect more words to correct.
Tajik is a separate option in the language list because it is written in Cyrillic. Choose Tajik for recordings from Tajikistan if you want Cyrillic output.
Classical poetry read aloud, such as Hafez or Rumi, uses older vocabulary and rhythm than everyday speech. It can be transcribed, but check recited verses against a printed text.
Persian script in the transcript
The transcript uses the Persian alphabet, including پ, چ, ژ and گ. Details of written Persian can vary from what you would type yourself.
- The half-space (zero-width non-joiner) in words like میخواهم may come out as a full space or with the parts joined together.
- Spoken Tehran forms such as میخوام can appear next to formal written forms in the same transcript.
- Now and then a character may appear in its Arabic form, such as ي instead of ی.
- English words may be kept in Latin letters or written in Persian script.
Language setting and code-switching
This page preselects Persian in the upload box. Auto-detect listens to the start of a file and can misjudge a short voice note, a clip that begins with music, or an interview that opens in English or German. With Persian selected, that guess is skipped. You can still change the language before uploading.
Persian speakers in the US, the UK, Germany and Sweden often mix in words from the local language. Short borrowed words usually come through. If a recording is mostly English with some Persian phrases, choose English for that file instead.
Who uses Persian transcription and what to do next
People bring oral-history recordings of parents and grandparents, research interviews, recorded poetry evenings and lectures, journalism interviews, podcasts and videos that need Persian subtitles. Some use it to prepare a draft of a client interview before a professional translator works on it. RocketScribe does not claim legal compliance, so have anything important checked by a person who speaks Persian.
TXT, DOCX, SRT, VTT and CSV keep the Persian text as written. The direct PDF option supports Western European characters only, so for Persian it opens a print-ready page that you save as a PDF from your browser. To read the content in English, use the Google Translate link on the transcript, or the AI prompt export to ask ChatGPT or Claude for a translation or summary. Speaker labels are available on paid plans or with a Minute Pack.