zen-scribe

Transcription

A speech-to-text transcription model supporting multiple languages with high accuracy. Handles diverse accents, noisy environments, and domain-specific terminology.

Specifications

Capabilities

• Multi-language speech-to-text
• Speaker diarization (who spoke when)
• Timestamp generation (word and segment level)
• Punctuation and capitalization
• Noise-robust transcription
• Domain-specific vocabulary support

Usage

API

from hanzoai import Hanzo

client = Hanzo(api_key="hk-your-api-key")

response = client.audio.transcriptions.create(
    model="zen-scribe",
    file=open("meeting.mp3", "rb"),
    language="en",
    response_format="verbose_json",
    timestamp_granularities=["word", "segment"],
)

for segment in response.segments:
    print(f"[{segment.start:.1f}s - {segment.end:.1f}s] {segment.text}")

See Also

• zen-translator -- Multi-language translation
• zen-dub -- Voice synthesis (text-to-speech)
• zen-live -- Real-time bidirectional translation