Use case
Client interviews, searchable in minutes
BrassTranscripts Legal turns client intake meetings, witness statements, and fact-finding calls into speaker-labeled transcripts paralegals and attorneys can search by participant and timestamp — replacing re-listening hours with text search seconds.
Create an accountRe-listening is the hidden cost of audio-only intake
A 60-minute client interview takes 60 minutes to review by audio playback — by transcript, the same content is searchable in seconds with timestamps that point back to the exact audio moment for context. BrassTranscripts Legal makes that conversion in 1–3 minutes per hour of recording.
Most paralegals replay client recordings looking for a specific commitment, statement, or claim. Transcripts make that workflow keyboard-driven instead: Cmd-F for the term, find the timestamp, click back to the audio if needed. A case with 40 hours of intake recordings becomes 40 hours of searchable text instead of 40 hours of re-listening.
Speaker identification keeps the conversation organized
BrassTranscripts Legal's automatic speaker identification distinguishes attorney from client, witness from interviewer, primary speaker from background voices — so transcripts read as structured dialogue rather than a single voice stream that requires manual annotation.
Each speaker turn gets a label (Speaker 1, Speaker 2, etc.) with timestamps. Replace the generic labels with names after download, or use the Speaker Name Assignment Helper prompt to do it programmatically based on voice characteristics in the first few exchanges.
From transcript to case insights
BrassTranscripts Legal includes AI prompts for thematic analysis, quote extraction, and action item tracking — prompts paralegals can run against transcripts in ChatGPT or Claude to surface case-relevant patterns across many client conversations without reading every line of every transcript.
- Thematic Analysis — surface recurring themes across client interviews
- Quote Extraction — pull specific statements with attribution and timestamps
- Action Item Tracker — extract commitments and follow-ups from a meeting
- Strategic Insights Extractor — identify case-relevant facts and decisions
See AI Prompts for the full toolkit.
Privileged communication considerations
BrassTranscripts Legal does not use customer audio or transcripts for AI model training, deletes audio files 24 hours after upload, and retains paid bulk transcripts for 10 days from payment before deletion — meaning no long-term copies of attorney-client privileged recordings remain on BrassTranscripts servers.
See Data Handling for the full picture: who has access, what's encrypted, what gets logged, and what is and isn't claimed about compliance.
Frequently asked questions
How private is a recorded client interview transcript?
BrassTranscripts Legal does not use customer audio or transcripts for AI model training, deletes audio 24 hours after upload, and retains paid transcripts for 10 days from payment before deletion. For attorney-client privileged interviews, treat the transcript file the same as any other privileged document in your matter management system.
How long can a client interview recording be?
Up to 2 hours per file, 250 MB maximum file size. For longer interviews (extended depositions of clients, multi-session intake), split the recording into separate files and upload as a batch — speaker labels remain consistent within each file but reset across files.
Will the AI capture client emotion and tone?
BrassTranscripts Legal produces text transcripts only — emotional tone, pauses, and demeanor are not annotated automatically. For witness assessment that depends on demeanor, the audio remains the source of truth; the transcript supports keyword search and AI prompt analysis but does not replace listening to the recording.
Can I run AI prompts on intake interviews?
Yes. The AI Prompt Toolkit includes prompts specifically for intake summaries, quote extraction with attribution, and thematic analysis. After downloading the TXT transcript, paste into ChatGPT, Claude, or your firm's preferred LLM with one of the prompts as the system instruction.
How accurate is speaker identification with phone-recorded interviews?
Phone interviews typically have lower audio fidelity than in-person recordings, but BrassTranscripts Legal speaker identification still distinguishes attorney from client when each speaks in clear turns. For multi-party calls (client + family member + interpreter), accuracy depends on whether each speaker is captured on a distinct channel or sufficiently distinct vocally.