Asylum interview recording to a full transcript with English under each paragraph
Upload the recording of an asylum or immigration interview in Pashto, Dari, Arabic, Tigrinya, Kurdish or another language and get every word written out as it was spoken, with a speaker label and a timestamp on each paragraph and an English translation underneath. It is an AI draft for checking what was said, not a certified transcript. A recording of up to 15 minutes is free in full; a longer one is $5, paid once, up to 6 hours.
- The applicant, the interpreter and the interviewer each written in the language they spoke
- A speaker label at the start of every paragraph, renamed on screen to Interviewer, Interpreter, Applicant
- English translation under each paragraph, the original language always first
- A timestamp like [1:42:07] on every paragraph, counted from the start of the file
- Up to 6 hours and 2 GB for one payment of $5
How to check an interview recording
-
1
Upload the audio
Drop the file here or paste a Google Drive or Dropbox link. Choose the applicant's language if you know it; English translation is already selected.
-
2
Read the free part
The first part appears on screen with labels, timestamps and English, so you can judge the quality before paying.
-
3
Unlock, rename, download
Pay $5 once, rename the speakers with Find and replace, and download Word, PDF, TXT or the CSV for your adviser.
Why the original words matter in an asylum interview
In an asylum interview the decision maker usually reads the interpreter's version, not the applicant's own words. If a date, a place name or a sentence about what happened was rendered differently, the written record can say something the applicant never said. Some authorities record interviews or hand over a copy on request, and some lawyers record with permission. Once you have that audio, the question is what was actually said, and in which language.
This tool writes each voice in the language it used. The applicant's Dari stays in Dari script, the interpreter's English stays in English, a Tigrinya answer is written in Ge'ez script, and a Kurmanji passage is written in Latin letters. Under each paragraph comes an English translation made by AI, so an adviser who does not speak the language can compare the applicant's words with what the interpreter said next.
What this transcript is, and what it is not
The transcript and the translation are made by an AI speech model. They are a draft to work from: a way to find the passages that need a closer look, not a certified transcript and not a sworn translation. If a lawyer wants to submit a passage, they may need a certified translator to confirm it from the audio. The timestamps make that quick, because the translator can go straight to [0:47:30] instead of listening to three hours.
Accuracy depends on the recording. Two people talking at once, a phone on the far side of the table, an echoing room or a strong regional accent all cause mistakes. Names of villages, people and dates are the words most worth checking by ear.
Languages people bring to this page
- Pashto and Dari from Afghanistan, both in Perso-Arabic script and kept apart
- Arabic in any dialect, Persian, Urdu and Somali
- Tigrinya and Amharic, written in Ge'ez script
- Kurdish: Sorani in Arabic script and Kurmanji in Latin letters, chosen as two separate languages
- Oromo or other languages not on the list may come out, but are not supported: check the free first part before paying
Labels for interviewer, interpreter and applicant
Every paragraph begins with a label such as "Speaker 1:". The recording is processed in parts of up to ten minutes, and on a long interview the same person can get a different number in a later part. Before downloading, use Find and replace on screen to turn each number into Interviewer, Interpreter or Applicant, and read the turns around each change to make sure they fit.
The downloads are Word, PDF and TXT, plus a CSV with one row per paragraph (File, Time, Speaker, Text, Translation) that opens in Excel. A lawyer can sort it by speaker or add a column for comments.
How the tool works, what it costs and how your files are handled is on the Interview Transcription page; prices for every tool are on the pricing page.
Last updated
Questions and answers
Is this a certified or official transcript of the interview?
No. It is an AI-made draft with an AI translation. It helps you find what to check; a lawyer may still need a certified translator for anything submitted.
Can ChatGPT transcribe and translate a 3-hour asylum interview?
Chat assistants can handle a short clip, but long audio often exceeds their upload limits, and a full transcript does not fit in one reply, so it tends to be shortened or stop partway. This tool writes every ten-minute part out in full and joins them in order.
What does it cost?
A recording of up to 15 minutes is free in full. A longer one shows its first part free and costs $5 once for the whole recording, up to 6 hours; Unlimited at $19 a week covers up to 12 hours per job and can be canceled any time.
Who can see the recording?
Only you. The file is sent to an AI model provider only to be processed and is not used to train models. Jobs started without an account are deleted after 7 days, and you can delete yours at any time.
The recording is in several files. Can they go together?
Yes. Upload the files of one interview in the same job; they are put in name order, so number them first if their names do not sort correctly.