Loading...

FTR Courtroom Transcription

Transcribe a .trm file without converting it first

AI drafts the first pass. You review, certify, and sign.

For transcribers and agencies handed a .trm file or an FTR case folder by a court, and for reporters who lose time exporting audio before they can start. Upload the .trm directly and get a draft with the speech attributed to the microphones in the room.

Pay as you go from $0.25/min. No subscription required.

Built for HIPAA workflows, with a BAA available · AES-256 encryption · No AI training on your data · First 30 minutes free

What FTR is, and what a case folder holds

Many U.S. courtrooms record on FTR (For The Record), a multi-channel system. Each microphone in the room, such as the bench, the witness stand, and the counsel tables, is recorded on its own channel inside one .trm file.

The recording usually arrives as part of a case folder. The .trm segments hold the multi-channel audio. Companion session and log files carry the order of the session and the case information. Keep the whole folder together as you received it, and upload the recording files as this page describes. The companion files help you confirm the order and details when you review.

A court session often arrives as several short recordings rather than one long file. Select all the clips from the hearing and upload them together. They are put in sequence using the times in their file names and returned as one transcript. If the order cannot be worked out with confidence, the upload is refused instead of guessed, because a transcript in the wrong order is a false record of what the court did. Up to 60 clips can be joined into one transcript.

Why a blended track fails in a courtroom

Flatten a courtroom into one audio track and several problems arrive at once. Microphones overlap. A quiet witness is buried under a loud bench microphone. Counsel and the judge speak at the same moment and the blended track cannot say which words belong to whom.

General diarization, which guesses speakers from voice characteristics alone, struggles in exactly these conditions. The FTR file already contains better evidence: which microphone each sound was strongest on. TranscribeLegal uses that.

The workflow, step by step

  1. Upload the .trm directly. There is no conversion step and no need to install FTR Player first.
  2. The microphone channels are read. The file is split into its individual microphone channels, and each is measured for speech activity.
  3. Silent channels are excluded. A microphone that carried no speech is left out of the draft, so you are not scrolling through empty tracks.
  4. Speech is attributed to the microphone it was strongest on. The channels that carried speech are combined and transcribed, and each passage is labeled with the microphone it was loudest on. Where two microphones were too close to call, such as simultaneous speech, the passage is labeled "Unidentified microphone" rather than assigned to a guess.
  5. Rename the labels. Microphones are labeled by hardware ("Microphone 1", "Microphone 2", and so on). Once you know who was on which microphone, rename them to the judge, the witness, or counsel.
  6. Review, certify, export. Review against the recording, certify, and export. The AI never certifies or signs a sworn record.

The labels are neutral on purpose. The FTR file does not say which microphone belongs to the judge, and the wiring differs from one courtroom to the next, so the draft will not invent a role it cannot know.

The honest limit: microphones are not people

The attribution tells apart microphones, not people. Two people sharing one microphone still arrive as one channel and one label, and the draft cannot separate them from each other by channel alone. Open microphones also pick up the room, so the judge is audible, faintly, on the witness microphone. That is why attribution is based on which microphone was strongest, and why close calls are labeled unidentified. Check those passages against the audio.

Two practical notes. First, FTR files are currently limited to about 55 minutes per file. For a longer session, upload the shorter segments the court exported, and join them as described above. Second, if a file will not open, the export may be incomplete or copied before the recording finished writing; re-export it from FTR. If the court protected the recording, ask the court for an authorised export. There is no workaround to promise here.

Manual route versus uploading the .trm

Export audio first, then transcribeUpload the .trm to TranscribeLegal
StartPrepare an audio export before any transcription beginsUpload the file as you received it
Room audioOften one blended trackMicrophone channels read; silent ones excluded
Speaker labelsAssigned by hand from the audioLabeled by microphone, which you rename
Draft timeAfter the export is doneTypically 10 to 30 minutes
CertificationYou certifyYou still review, certify, and sign

Cost example: a 45-minute FTR segment is $11.25 on Pay As You Go ($0.25/min). Because FTR files are currently limited to about 55 minutes each, a three-hour court session is uploaded as several segments, and the total is still $0.25 per minute of audio with the $5 minimum applying to any file under 20 minutes.

How it works

Step 1 — Upload

Drag and drop any audio or video file: MP3, WAV, MP4, MOV, and more, including FTR .trm courtroom files.

Step 2 — Transcribe

AI processes the recording in minutes. It identifies speakers, adds clickable timestamps, and recognises legal terminology.

Step 3 — Review and export

Search and edit in the browser, export the Q&A-format RTF for Word, plain text, SRT/VTT, or JSON — then certify and export a certified PDF once you have signed off.

Pricing

Pay per minute or save with a monthly plan. No contracts, no per-page fees, no rush surcharges. Every plan includes speaker identification, timestamps, and all export formats. First 30 minutes free, no credit card required.

PlanPriceIncludedOverage / notes
Pay As You Go$0.25/min—$5 minimum per file, no monthly fee
Basic$49/mo300 min$0.20/min overage
Standard$99/mo750 min$0.18/min overage, priority processing
Advanced$249/mo2,500 min$0.15/min overage, API access, BAA included

Annual billing saves 2 months: Basic $490/yr, Standard $990/yr, Advanced $2,490/yr. Pricing is per minute of audio and is the same nationwide. See full legal transcription pricing or the legal transcription pricing guide.

Security for sensitive recordings

  • Encrypted in transit (TLS 1.3) and at rest (AES-256).
  • Built for HIPAA workflows, with a BAA available.
  • No AI training on customer data.

The certificate of accuracy is part of the human certification workflow the reviewer completes. The raw AI draft is never described or delivered as certified. Read more on the security page or see the Business Associate Agreement.

Frequently asked questions

Upload a .trm file and see the draft

Try an FTR recording on your first 30 free minutes. No credit card, no conversion step, and the review and certification stay with you.

Running a law firm? Transcription is built into ProvaLens, the AI document-intelligence platform for firms — transcription minutes included, plus AI search, case-timeline generation, and contradiction detection. ProvaLens starts at $249/mo. See ProvaLens.