docs(ML-93): answering machine detection for live STT

[jqueguiner]

What

Documents the answering_machine message for live STT.

• chapters/live-stt/features/answering-machine-detection.mdx — the feature page
• api-reference/v2/live/callback/answering-machine.mdx — payload reference, same shape as the existing speech-start / speech-end pages
• asyncapi.yaml — AnsweringMachineMessage schema with examples, the answeringMachine channel message, the onAnsweringMachine operation
• docs.json — both pages in the navigation

The message

{
  "session_id": "550e8400-e29b-41d4-a716-446655440000",
  "created_at": "2026-09-09T12:34:11Z",
  "type": "answering_machine",
  "data": { "kind": "machine", "confidence": 0.78, "time": 5.24, "channel": 0 }
}

Three things stated explicitly, because all three are easy to get wrong

confidence is a ranking score, not a probability. 0.78 does not mean "78% chance of a machine". The page says so in a <Warning> and tells integrators to tune a threshold on their own traffic instead.

Silence is reported as human. A callee who never speaks produces no speech at all, and staying on the line beats hanging up on someone quiet.

time is the audio position, not wall clock. It shares its reference with speech_start, speech_end and transcript, so messages can be ordered against each other. A client pushing audio faster than real time will see time: 5.24 arrive within a fraction of a second, so created_at is the field to correlate with their own logs.

All three would otherwise surface as support tickets.

Verified

asyncapi.yaml parses and the new nodes resolve on the liveTranscription channel. docs.json is valid JSON with both entries in the navigation. Mintlify validation and link-rot checks pass on the branch.

Refs ML-93

[mintlify]

Preview deployment for your docs. Learn more about Mintlify Previews.

Project	Status	Preview	Updated
gladia	🟢 Ready	View Preview	Sep 9, 2026, 4:23 PM

💡 Tip: Enable Automations to automatically generate PRs for you.