Skip to content

Docs

Build with Turkish speech

Two Turkish speech models, one API key. Alania turns text into speech; Duyu turns speech into text. Both work with the OpenAI client libraries: change base_url and you are done.

QuickstartParameters reference
First request
curl https://voice.patientdesk.ai/v1/audio/speech \
  -H "Authorization: Bearer pd_live_..." \
  -H "Content-Type: application/json" \
  -d '{ "model": "alania-v2",
        "voice": "alania2-f1",
        "input": "Talebinizi ilgili ekibe aktardım, sonucu size kısa mesajla ileteceğiz." }' \
  --output talep.wav
Result: talep.wav

Products

Guides and reference for each product.

Text to speech

Alania

Built-in voices, style instructions, voice design from a description, consented cloning. 48 kHz audio that streams as it is made.

Speech to text

Duyu

File transcription and live streaming. Compatible with OpenAI Whisper clients; word timestamps, SRT and VTT.

Models

The ids you put in model.

alania-v2DefaultText to speech. Built-in voices, style, design, cloning.48 kHzantalia-miniFreeThe open-source Antalia-2 Mini. One voice, no quota used.48 kHzalania-v1LegacyThe first Alania. One voice, no extra fields.24 kHzduyu-1STTSpeech to text: files and live streams.REST · WebSocket

Examples

End-to-end projects you can copy and run.

Speech in a web appKeep your key on the server: a small route forwards the request to Alania and streams the audio back, and the browser starts playing on the first chunk.A phone line with TwilioAnswer a call, transcribe the caller live with Duyu and reply in Alania's voice. Audio stays 8 kHz μ-law both ways; nothing to resample.Live captions in the browserStream the microphone to Duyu and show the words as they are spoken: partial drafts while someone talks, a settled final line when they pause.Narrating a long textTurn an article or a book chapter into one audio file: split the text into 5,000-character parts, ask for raw PCM, and write a single WAV.Subtitles for a videoPull the audio track out with ffmpeg, ask Duyu for srt, and attach the subtitles to the video.

Base URLs

Every request goes over HTTPS or WSS; errors come back as JSON in the OpenAI shape.

Alania · REST

POST/v1/audio/speech

Alania · WebSocket

WSS/v1/audio/speech/stream

Duyu · REST

POST/v1/audio/transcriptions

Duyu · WebSocket

WSS/v1/audio/stream