Audelyr
Contact salesStart building

Developer documentation

Production + Preview

Audelyr developer documentation

Build with realtime speech-to-text, speech translation and speech synthesis APIs. Pure text translation is not a product capability.

Open the console ↗Read the docs

Capabilities

One interface.
Clear boundaries.

Designed as composable infrastructure, with Preview labels wherever production behavior is not yet available.

01

Realtime STT

Documented, observable and designed to fit into your existing product architecture.

02

Speech translation

Documented, observable and designed to fit into your existing product architecture.

03

TTS

Documented, observable and designed to fit into your existing product architecture.

04

Async transcription

Documented, observable and designed to fit into your existing product architecture.

Interface example

Start with a request.
Inspect every field.

Use a project API key from your server and follow the capability status documented on this page.

curl https://api.audelyr.ai/v1/realtime/sessions \
  -H "Authorization: Bearer $AUDELYR_API_KEY" \
  -H "Idempotency-Key: realtime-demo-001" \
  -H "Content-Type: application/json" \
  -d '{"audio":{"format":"pcm_s16le","sample_rate_hz":16000,"channels":1}}'
01

Overview

Build with realtime speech-to-text, speech translation and speech synthesis APIs. Pure text translation is not a product capability.

02

Before you begin

Create a project API key in the console and keep it on your server.

03

Which URL does what

api.audelyr.ai is the production REST API (Bearer aud_live_ key). realtime.audelyr.ai is the production WebSocket audio gateway (short-lived Audelyr Session Token). The Console reaches its private control plane through same-origin /api/platform routes with a Firebase ID Token; there is no public Control API hostname. api.audelyr.ai/v1/preview remains the migration-only Preview surface (aud_preview_ key).

04

Request lifecycle

Realtime connections return events as audio is processed. HTTP and SSE endpoints return structured errors when a request cannot run.

05

Availability

Each guide states whether its capability is available today or still planned. Do not send production traffic to an endpoint marked unavailable.

Early Access

Build the interface
before the infrastructure.

Explore the honest Preview, then tell us what production voice needs from Audelyr.

Create an accountContact sales

Documentation search

Find an interface,
guide or reference.

/docsAudelyr developer documentation

Build with realtime speech-to-text, speech translation and speech synthesis APIs. Pure text translation is not a product capability.

/docs/quickstartQuickstart

Create a project API key, then create a realtime session or call an HTTP endpoint from your server.

/docs/authenticationAuthentication

Use project API keys on your server. Realtime clients use short-lived session tokens issued by your backend.

/docs/realtime-apiRealtime Speech API

Create a session with POST /v1/realtime/sessions, then stream microphone audio over WebSocket to /v1/realtime.

/docs/realtime-translationRealtime voice translation

Set translation.mode to one_way or bidirectional when starting a realtime WebSocket session. Receive transcript.final and translation.final events in the same connection.

/docs/speech-to-textRealtime Speech-to-Text

Not verified in this deployment: the Realtime WebSocket API processes audio as it arrives and can retain final transcript segments per session.

/docs/text-to-speechText-to-Speech

Not verified in this deployment: POST /v1/speech/synthesize returns generated speech and POST /v1/speech/streams provides streaming SSE output.

/docs/translationSpeech-to-text translation

Not verified in this deployment: configure one-way or bidirectional audio translation and receive original and translated tokens; pure text translation is not offered.