Skip to content

Free preview

Legal Audio Transcription

Drop a recording of a hearing, deposition, client call or dictated memo and read the first 5 minutes as text with timestamps and speaker labels. No account; the file is deleted once the transcript is delivered.

  • Free preview: first 5 minutes
  • Deleted after processing
  • No signup

Drop a deposition, hearing or client-call recording here

or

Supports MP3, M4A, WAV, OGG/OPUS, AAC, FLAC and WEBM

How it works

Legal Audio Transcription in three steps

01

Drop the recording

Hearing audio, a recorded call or a dictated memo. The browser keeps the first 5 minutes.

02

Timestamps and speakers

Every turn is stamped and labelled. Language detected automatically.

03

Review and cite

Copy the text or download a .txt. Full recordings, summaries and a matter library are in the app.

Hours of audio, searched in seconds

A two-hour deposition, a recorded client intake, a hearing you could not attend: the words are all there, but finding the one statement that matters means scrubbing through audio. As text, with a timestamp on every turn, the same recording is searchable, quotable and citable to the second. This preview shows you that on the first 5 minutes of your own file.

What you get, and what it is not

Every speaker turn carries a timestamp and a label — Speaker A, Speaker B — consistent through the transcript, so a witness and an examiner stay apart. Punctuation is added and fillers are lightly cleaned for readability. It is a working transcript for review, discovery and drafting. It is not a certified transcript: courts that require certification require a human transcriptionist's attestation, and nothing produced here carries one.

What the preview does with the recording

Your browser reads the file and keeps the first 5 minutes, cutting a longer recording locally and re-encoding that part as 16 kHz mono WAV so the rest is never uploaded. The audio goes over an encrypted connection to the transcription engine behind the SozAI app (AssemblyAI). The language is detected automatically across 99 languages. When the text is on your screen, the audio and transcript are deleted at the provider; this site stores neither.

Confidentiality

This is a public preview, not a service under an engagement or data-processing agreement, and audio is processed by a third-party provider in the US region unless the site has been switched to its EU endpoint. Use it with recordings you are free to process that way — a public hearing, a de-identified excerpt, your own dictated memo — and keep privileged material for tools your firm has vetted. The SozAI app's storage and encryption are described on the security page.

Formats and limits

  • M4A, MP3, WAV, AAC, FLAC, OGG/OPUS, WEBM, up to 25 MB as-is; longer files are trimmed to 5 minutes in the browser first (up to 30 MB on a computer, 12 MB on a phone).
  • Three previews per device per day, no account.
  • Courtroom recordings from a fixed microphone transcribe well; a phone in a pocket during a site visit does not.

Full recordings

The SozAI app transcribes any length, keeps timestamps and speaker labels, produces a summary, and stores every recording in a library you can search across matters. The first 30 minutes are free.

Answers

Legal transcription — frequently asked questions

Is this a certified transcript?

No. Certified transcripts require a human transcriptionist's attestation. This produces an accurate working transcript with timestamps for review, search and drafting — not a document to file where certification is required.

Can I use it for privileged recordings?

The preview is a public tool processed by a third-party provider without an engagement or data-processing agreement, so treat it as you would any unvetted service: public hearings, de-identified excerpts and your own dictation, not privileged material.

Does it timestamp every statement?

Every speaker turn carries a start time and a speaker label, consistent through the transcript. That is enough to cite a statement to the second and to find it again in the audio.

How are multiple speakers handled?

Voices are separated and labelled Speaker A, Speaker B and so on. Names are not inferred; you assign them. Examiner and witness, or two parties on a call, stay apart as long as both are audible.

What languages does it support?

The language is detected automatically across 99 languages. A hearing conducted through an interpreter comes back in whichever language dominates the recording.

How long a recording can I preview?

Any length, but you get the first 5 minutes. Files up to 25 MB are sent whole; longer recordings are cut to 5 minutes in your browser before upload. A two-hour deposition is a job for the app.

Is the audio retained?

No. It is transcribed over an encrypted connection and deleted at the provider together with the transcript once the result is delivered. This site stores neither.

How accurate is it on courtroom audio?

Fixed-microphone recordings with speakers taking turns transcribe close to verbatim. Crosstalk, a distant gallery and legal Latin spoken quickly cost words. Review anything you intend to quote.

The whole recording, not just the first minutes

Record or import any length in the app: speaker labels, timestamps, a summary and translation into 100+ languages.

Get the App — Free

Free on iOS and Android. 30 minutes of transcription included, no card.