Your audio never leaves your device
Transcription for lawyers that keeps client audio on your device
Turn client interviews, witness statements, and recorded calls into text without sending the audio to a cloud service. Everything runs in your browser, so the recording stays under your control.
Takes a little extra time, and your audio still never leaves your device. Labels go by how voices sound, so treat them as a best effort: one person can be split across two or three labels, similar voices can be merged, and someone joining partway through may be folded into a speaker already talking. Giving a number usually helps. Check them before relying on them.
Specifying this number improves accuracy. Count people who really take part, whenever they join. Someone who only says a word or two is better left out.
Timestamps each word instead of each phrase, so short interjections like "Yeah" go to the right person. On a busy four-person recording this corrected about 1 word in 5; it changes little when people talk in long turns. Adds roughly 40% to the time on a computer, and almost nothing on a phone.
Phone audio loses the high frequencies that separate consonants, so they are put back before transcribing. Only phone-quality recordings are treated, and you are told when it happens. Also even out volume is for one case: someone far from the microphone and much quieter throughout. It cost accuracy on other recordings, so use it only if a quiet person came out badly.
Worth filling in. The speech model knows ordinary English, not your subject, so a word it has never heard becomes whatever sounds closest, and it usually gets that word wrong every single time it comes up. A sentence is enough, and the names and jargon in it are picked out and watched for. It cannot guess a word you do not mention, so name the people and the terms that matter.
Free, no signup. The first time you use it, it takes a minute or two to set up. After that setup is quick. Works with MP3, WAV, M4A, OGG, FLAC and AMR, and long recordings are handled in parts.
Why on-device transcription fits legal work
A privileged recording is not something you want to hand to a third-party server. Once audio is uploaded, you are trusting another company's security, retention schedule, subprocessor list, and data practices with material that may be privileged or subject to a protective order. You are also creating a copy that exists outside your file, in a place you would have to account for if anyone ever asked where the recording had been.
MeetingScribe avoids that by never uploading the file. The speech model runs in your browser, the recording is transcribed on your machine, and the text stays with you. There is no vendor copy because there is no vendor step.
Where a working transcript helps, and where it does not
This distinction matters more in legal work than anywhere else, so it is worth being blunt about it. What this produces is a working transcript: fast, searchable, good enough to review, quote from with verification, and build a chronology out of. What it is not is a certified transcript.
Anything destined for the record still needs a certified court reporter or transcriptionist under the rules that apply in your jurisdiction. No automated system, on-device or cloud, and regardless of vendor claims, substitutes for that. Treat the output the way you would treat a very fast paralegal's first pass: useful immediately, verified before it matters.
Realistic uses:
- Intake and client interviews. Get a searchable record of a first meeting without splitting your attention between listening and typing.
- Witness statements. Produce a draft you can read back and correct with the witness, rather than reconstructing from memory afterwards.
- Recorded calls. Search a call for what was actually said rather than scrubbing audio for the one exchange you remember.
- Reviewing produced audio. Turn recordings received in discovery into text you can search, index, and cite by timestamp.
- Preparation. Build a timeline from your own recorded notes and dictation.
A workflow that survives contact with a real file
- Split long audio first. Export a deposition or long interview into forty minute parts named matter_part1, matter_part2, and so on. Selecting them together joins them into one transcript in the right order.
- Turn timestamps on. A transcript without timestamps is hard to verify against the audio later. With them, every passage you might quote points back at the moment it came from.
- Use the small model for phone and speakerphone audio. The larger download pays for itself on the recordings legal work actually produces, which are rarely studio-quality.
- Rename speakers before you export. Replacing SPEAKER_00 with the actual names once, in the page, is faster than a find and replace in your document later.
- Correct proper nouns in the editor. Speech models mishear names, case citations, and technical terms more than ordinary words. Fix them while the audio is fresh.
- Save into the matter file, not a downloads folder. The transcript is a document with the same sensitivity as the recording, and it deserves the same handling.
Consent is your responsibility
Laws about recording conversations differ widely. Some jurisdictions permit one-party consent, others require every participant to agree, and the applicable rule can depend on where each person was located rather than where you were. Cross-border calls complicate this further.
MeetingScribe is a tool for turning audio you already have the right to use into text. It does not advise you on whether a particular recording was lawfully made, and it cannot tell whether the file you dropped in was obtained properly. Our plain-English guide to recording law is general background only and should not be relied on for a specific matter. You are far better placed than we are to answer this question for your own jurisdiction.
Retention, and the copy you did not think about
On-device transcription removes the vendor copy. It does not remove the copies you make yourself, and those are the ones that tend to go unmanaged. The downloaded transcript, the subtitle file, the version pasted into an email, and the original recording on the phone that captured it are all now discoverable material sitting in your own systems.
One further caution about the notes feature: after transcribing, the tool can copy a prompt containing your transcript to your clipboard so you can paste it into an AI chat. Nothing is sent by the copy itself, but pasting privileged content into a consumer AI service is a decision with its own consequences, and it is one to make deliberately rather than by habit. For privileged material, writing the summary yourself from the transcript is the conservative choice.
Nothing on this page is legal advice, and using this tool creates no attorney-client or other professional relationship. It is a description of software written for lawyers who already know their own obligations.
Last reviewed: August 2026
Frequently asked questions
- Can I use this as a deposition or hearing transcript?
- No. This produces a working transcript, not a certified one. Depositions, hearings, and anything intended for the record require a certified court reporter or transcriptionist under the applicable rules. Use this for your own review, preparation, and file notes.
- Can I record a client or a call and transcribe it?
- You are responsible for consent. Recording and consent laws vary by state and country, and some require every party to agree. Confirm the rules that apply where each participant is located before recording. This page is general information, not legal advice.
- Is this HIPAA or otherwise compliant?
- We make no compliance claims and offer no business associate agreement or data processing agreement. The design keeps audio on your device, which is why it suits confidential work, but your obligations remain yours to assess.