29Court voice pipeline
Ingest
Speak, type, or drop in audio. Then watch your words travel the pipeline, stage by stage, with real output at every step.
New bark
Hold the button and speak. Your browser transcribes on the spot. Audio never leaves this device.
Type below, then send. Rambling is fine. The pipeline sorts it out.
Audio uploads are not transcribed server side yet. The file stays on this device. Type what it says in the transcript box, and that is what the pipeline processes.
Structured note
museAction items
Entities
Buckets
Cleaned transcript
Pipeline, live
-
Received
Waiting for your first bark.
-
Transcribe
Transcript passes through.
-
Clef classification
Typed decision, real model.
-
Rule sets
Buckets, cues, entities.
-
Route
Destination per client config.
-
Done
Note lands in the inbox.
Inbox
No barks yet
Send your first one above and it will land here, structured and routed.