Deposition Transcription — Working Drafts From $0.01/Minute

Not as the official record. AI transcripts are working drafts, not certified or sworn. They can be admissible as demonstrative aids or working copies when both sides stipulate, but anything filed, sworn, or contested needs a certified court reporter or licensed transcriptionist. Standard workflow: AI draft for prep, summary, and quote-pulling; certified transcript for the record.

For clean recordings of native English speakers in a quiet setting, modern AI transcription delivers 92–97% word accuracy. Accuracy drops with heavy accents, overlapping speakers, poor audio, and specialized legal terminology. Always review the AI draft against the recording before quoting from it. Speaker labels and word-level timestamps make verification fast — click any line to jump back to the source second.

VexaScribe diarization handles up to 50 distinct speakers per file (best accuracy with 2–6). Typical depositions have 4–6 voices: deponent, examining attorney, defending attorney, court reporter, sometimes co-counsel or an interpreter. The AI assigns generic labels (Speaker 1, Speaker 2, ...) which you rename once in the editor. The renames propagate across the entire transcript.

Typically 5–15 minutes of processing time for a 4-hour audio file on standard plans. Upload, wait, download. Compare: human services like Rev Legal or GoTranscript take 12–48 hours for the same file; certified court reporters take days to weeks for a formal transcript. When you need a draft to prep tomorrow, AI is the only option that fits.

MP4, MOV, MKV, WEBM (video), and MP3, WAV, M4A, FLAC, OGG (audio). Most video deposition platforms (Zoom, Microsoft Teams, OnCue, RemoteDepo, Veritext Virtual, Esquire Connect) export to MP4 — that uploads directly with no conversion. Max upload size is 5 GB per file on standard plans.

All uploads are encrypted in transit (TLS 1.3) and at rest (AES-256). No human reviewer ever listens to your audio — transcription is fully automated. Only your account can access your files. Configure auto-delete per upload: 24 hours, 7 days, 30 days, or never. EU data residency available on request. We are not a HIPAA business associate — if your deposition includes PHI, plan accordingly or use a HIPAA-attested vendor for that file.

Verbatim mode preserves filler words ("um", "uh"), false starts, interruptions, and exact phrasing — useful when word-for-word accuracy matters, like deposition drafts where the exact hedge or pause might be legally significant. Standard mode lightly cleans the transcript for readability. For depositions and witness testimony, default to verbatim.

Yes — for the recording afterward. Most remote depositions are recorded by the platform (Zoom, Teams, RemoteDepo, Veritext Virtual). Export the MP4, upload to VexaScribe, get a draft in minutes. For the live official record, the court reporter on the deposition handles certification; AI is the back-end draft layer.

A court reporter (stenographer) captures spoken testimony live in real time using stenography or voice-writing, certifies the transcript as accurate, and produces a sworn paginated record for the case file — typical cost $4–$10 per minute equivalent plus per-page production fees. Deposition transcription (in the AI sense) converts an already-recorded deposition into searchable text after the fact — typical cost $0.003–$0.01 per minute for a working draft. They are complementary, not competing.

A 4-hour deposition (240 minutes) costs: certified court reporter $960–$2,400+ (depending on rate and page fees); Rev Legal human transcription $300; GoTranscript $360; VexaScribe Pro plan effective ~$1.00 ($10/month for 2,500 min); VexaScribe Studio effective ~$0.80 ($20/month for 6,000 min). AI is 300–3,000× cheaper than a court reporter for the same length — the trade is certification vs working draft.