Enables diarization

Uses keyterms

Uses scribe_v2 model

Groups output by speaker

Accesses word.speaker_id

Word-level timestamps

Uses word.start and word.end

Filters by word type

Enables audio event tagging

Uses useScribe hook

Token-based auth

Handles both transcript types

Microphone config

Manual commit strategy

PCM 16-bit format

Correct chunk size

Base64 encoding

Explicit commit call

VAD enabled

VAD threshold tuning

Partial for display

Committed as source of truth

Uses scribe_v2_realtime

Entity detection enabled

Accesses entities in response

Language code provided

Uses scribe_v2 model

Reads language probability

Reads detected language code

Correct JS package

Entity output included

additional_formats used

SRT format requested

with_raw_response used

request-id captured

response.parse() called

Uses scribe_v2 model

Request ID in output

Status checks both states

VAD commit strategy

CommitStrategy imported from @elevenlabs/react

previous_text field used

previous_text length constraint

Uses scribe_v2_realtime model

Token-based auth

Character-level timestamps

Diarization enabled

num_speakers set to 3

Uses scribe_v2 model

subtitle_data.json written

Speaker ID in output

Correct Python package

Uses cloud_storage_url

Enables multichannel

Uses scribe_v2 model

Correct Python package

Per-channel output

JSON file written

No local file download

no_verbatim enabled

Seed parameter used

Uses scribe_v2 model

transcript.txt written

transcript_meta.json written

Language fields in metadata

Correct Python package