Our standard
The record as spoken, in the format the court expects.
A transcript is evidence of what was said. Ours is produced to four rules: two passes against the audio, verification of every proper noun and citation, formatting to the governing order, and a written account of anything we could not resolve.
- 01
Two passes against the audio
The transcriptionist listens to the entire recording while producing the transcript. The reviewer listens again while checking it. Reading without audio catches punctuation; only listening catches a dropped 'not'.
- 02
Verified, or marked
Parties, counsel, witnesses, firms, places, products, case citations and statutes are confirmed against a valid source: the case paperwork first, then official rosters, bar directories and court records. A name that cannot be confirmed after a real search is marked (phonetic) in the transcript and listed in the delivery note. We do not pick the most likely spelling.
- 03
Formatted to the governing order
Court transcripts follow the court's format order: caption wording, appearance page, speaker designations, examination headers, exhibit parentheticals, lines per page, certificate wording. Depositions follow your agency's or firm's template. Where there is no template, ours is a 25-line pleading format.
- 04
Verbatim, at the level ordered
Strict verbatim keeps every false start, repetition and filler. Clean verbatim removes an attorney's or judge's false starts and fillers, and nothing of a witness's. You choose the level on the order. Testimony is never improved.
| What you will see | |
|---|---|
| Words that cannot be made out | (indiscernible), with the timestamp in the delivery note |
| Spelling that cannot be confirmed | the name followed by (phonetic) |
| Speaker who cannot be identified | a consistent label such as UNIDENTIFIED SPEAKER or MALE VOICE 1, never a guess at a name |
| Interpreter on the record | the interpreted testimony attributed to the witness; the interpreter's own remarks attributed to the interpreter, per the tribunal's rule |
| Sidebar, bench conference, off-record discussion | a parenthetical, with times where the format order requires them |
| Recording played in the proceeding | a parenthetical noting the playback; the recording's content is not transcribed unless ordered |
| Non-English passage | noted with the time and the language; not translated |
| Paperwork that conflicts with the record (a caption spelling, a missing appearance) | the transcript follows the record; the conflict is listed in the delivery note for you to resolve from the file |
Speech recognition, stated plainly
A first draft may be produced by speech-recognition software running on a business service whose terms prohibit retaining or training on your audio. Every word of that draft is then worked against the recording by a legal specialist and checked by a second. Client audio is never uploaded to consumer AI services or public transcription tools.
How the standard is kept
Every transcript is rated by its reviewer. Specialists begin on supervised work with full-audio review and a redline on every job, and move to self-proofing only on measured error rates. The reviewer's redline is available on request for any job.
What we never do
- We never guess a name.
- We never clean up a witness.
- We never reconcile the caption and the record silently.
- We never run client audio through consumer AI or public transcription tools.
- We never deliver without the second pass.
Ready for litigation-ready transcripts you can rely on?
See the standard on the page: send fifteen minutes and judge it on your own recording.