Overview

Where every model and voice stands, from raw recordings to production.

TTFA gate ≤ 200ms · p95
—
Intake
loading…
—
Training runs
loading…
—
Voices
loading…
—
Delivery
loading…
Checking the API…

How this connects

One pipeline, not separate tools — an intake batch becomes a training run, a run produces a voice, a voice ships through delivery, and production feeds back into the next run.

  1. recordings + consent
  2. Intake
  3. Training runs
  4. Voices
  5. Delivery
  6. Observability

Onboard a new voice

Start with a client's recordings, or open a call-capture window.

+ New intake

See the product

Run a simulated support call end to end — the STT → agent → TTS loop.

▶ Start a call