Azure Speech TTS Provider for OpenClaw

Azure Speech is a text-to-speech provider bundled with Azure AI Speech. OpenClaw communicates directly with the Azure Speech REST API using SSML, producing MP3 for standard replies, native Ogg/Opus for voice notes, and 8 kHz mulaw for telephony channels like Voice Call. The request includes the provider-owned output format in the X-Microsoft-OutputFormat header.

Detail	Value
Provider ID	`azure-speech` (alias: `azure`)
Website	Azure AI Speech
Docs	Speech REST text-to-speech
Auth	`AZURE_SPEECH_KEY` plus `AZURE_SPEECH_REGION`
Default voice	`en-US-JennyNeural`
Default file output	`audio-24khz-48kbitrate-mono-mp3`
Default voice-note file	`ogg-24khz-16bit-mono-opus`

Getting started

Create an Azure Speech resource

From the Azure portal, provision a Speech resource. Copy KEY 1 found under Resource Management > Keys and Endpoint, and note the resource location, for instance eastus.

AZURE_SPEECH_KEY=<speech-resource-key>
AZURE_SPEECH_REGION=eastus

Select Azure Speech in tts

{
  tts: {
    auto: "always",
    provider: "azure-speech",
    providers: {
      "azure-speech": {
        voice: "en-US-JennyNeural",
        lang: "en-US",
      },
    },
  },
}

Send a message

Transmit a reply over any connected channel. OpenClaw generates audio through Azure Speech, returning MP3 for standard audio or Ogg/Opus when the channel anticipates a voice note.

Configuration options

All configuration options are located under tts.providers["azure-speech"].

Option	Description
`apiKey`	Key for the Azure Speech resource. Defaults to `AZURE_SPEECH_KEY`, `AZURE_SPEECH_API_KEY`, or `SPEECH_KEY`.
`region`	Region for the Azure Speech resource. Defaults to `AZURE_SPEECH_REGION` or `SPEECH_REGION`.
`endpoint`	Optional override for the Azure Speech endpoint. Falls back to trusted `AZURE_SPEECH_ENDPOINT`.
`baseUrl`	Optional override for the Azure Speech base URL.
`voice`	Azure voice ShortName (default `en-US-JennyNeural`). Legacy alias: `voiceId`.
`lang`	Language code for SSML (default `en-US`).
`outputFormat`	Format for audio file output (default `audio-24khz-48kbitrate-mono-mp3`).
`voiceNoteOutputFormat`	Format for voice-note output (default `ogg-24khz-16bit-mono-opus`).
`timeoutMs`	Timeout override for requests in milliseconds. Falls back to the global `tts.timeoutMs`.

The provider is considered configured when apiKey is set along with one of region, endpoint, or baseUrl. Environment variables serve only as fallbacks for config keys that remain unspecified. Workspace .env files cannot define AZURE_SPEECH_ENDPOINT; use the process environment, global runtime dotenv, or explicit configuration for endpoint routing.

Notes

Authentication

Azure Speech requires a Speech resource key, not an Azure OpenAI key. This key is transmitted as Ocp-Apim-Subscription-Key; OpenClaw derives https://<region>.tts.speech.microsoft.com from region unless endpoint or baseUrl is supplied.

Voice names

Use the Azure Speech voice ShortName value, for example en-US-JennyNeural. The bundled provider can enumerate voices through the same Speech resource, excluding those marked as deprecated, retired, or disabled.

Audio outputs

Azure supports output formats like audio-24khz-48kbitrate-mono-mp3, ogg-24khz-16bit-mono-opus, and riff-24khz-16bit-mono-pcm. For voice-note targets, OpenClaw requests Ogg/Opus so channels can send voice bubbles natively without a separate MP3 conversion, and it enforces raw-8khz-8bit-mono-mulaw when telephony targets are used.

Alias

Existing configurations can use azure as a provider alias, though azure-speech should be used in new configurations to prevent confusion with Azure OpenAI model providers.

Text-to-speech, A TTS overview, provider details, and tts configuration.
Configuration, A full configuration reference that covers tts settings.
Providers, Every bundled OpenClaw provider.
Troubleshooting, Frequent issues and steps for debugging.