OpenAI-compatible transcription API from $0.06 an hour

Same request, same response

Set the SDK’s base URL to https://api.avencera.com/v1. Requests go to /audio/transcriptions as they do at OpenAI, and responses come back as text, json, verbose_json, srt or vtt, with word or segment timestamps.

Speaker labels in the same call

Set speaker_labels to true and every segment carries its speaker, for $0.07 an hour or more. Ask for diarized_json to get the segments the OpenAI SDK types for speakers.

Open models, priced by the hour

NVIDIA Parakeet TDT 0.6B v2 transcribes English for $0.06 an hour of audio, and v3 covers 25 European languages for $0.07.

OpenAI SDK requests take files up to 2 hours and 100 MB. Longer recordings, up to 24 hours, go through a job, not cut into short clips first.

A speaker transcript in Python

from openai import OpenAI

client = OpenAI(base_url="https://api.avencera.com/v1", api_key="YOUR_AVENCERA_KEY")

with open("meeting.mp3", "rb") as audio:
    transcript = client.audio.transcriptions.create(
        model="avencera/parakeet-tdt-0.6b-v2",
        file=audio,
        extra_body={"speaker_labels": True},
    )

Questions

Do the official OpenAI SDKs work with Avencera?

Yes. Point the SDK’s base URL at Avencera, use your Avencera API key, and change the model name. The request and response stay the same.

Do I need a subscription?

No. Buy prepaid credit and pay only for the audio you send. Credit never expires.

Which model should I use?

Parakeet v2 for English and v3 for other European languages. The model chooser recommends one for each kind of audio.