Proof against the audio.

A scoping player built for proofing, and a second listen from three independent speech recognizers on the passages that need it.

Illustration: ripples spreading across still water from one drop

A scoping player for long audio

  1. Dead air plays faster

    Silence is sped up, not cut, with sliders for playback speed and pause length.

  2. Keys and pedals

    Space starts and stops playback, and the arrow keys skip five seconds. USB foot pedals that send ordinary keys work.

  3. Your notes stay in charge

    Your transcript stays the working text. Nothing replaces it unless you choose.

Three recognizers, one decision

Each window of audio is transcribed by three independent speech recognizers. Where they disagree with your notes, you hear a short clip and decide.

  1. You choose

    Keep your notes, take a reading, type your own text, or mark it unclear.

  2. You can overrule anything

    Some passages are settled by rules when the listeners agree. You can change any of those decisions.

  3. Back into your file

    Your decisions are written back into the file you uploaded, so it goes back to your editor.

Where the audio goes

Proofing sends audio to third-party speech recognition services for the second listen. Proofing audio and its review notes are deleted after 30 days by default.

Then build the final

Export the proofed text and build the transcript in the state's format.

Transcripts

Reporter Desk is in early access

There is no pricing or self sign-up yet. Tell us how you work and which states you report in, and we will write when you can onboard.