Azure AI Speech text-to-speech vs Rime TTS API + MCP

Azure AI Speech text-to-speech

Real-time synthesis keeps neither the input text nor the output audio. An Azure subscription needs a card, even for the free F0 tier.

Rime TTS API + MCP

Only character counts are kept by default, and customer data isn't used for training without an opt-in. No OpenAPI or AsyncAPI file and no official SDKs.

Azure AI Speech text-to-speech

1. Send SSML with <speak> and <voice>, and set X-Microsoft-OutputFormat and User-Agent.
2. On 429 retry with backoff, and try the voice's home region or another region rather than asking for more quota.
3. Keep each real-time request under 10 minutes of audio, or use batch synthesis.
4. Use Entra ID tokens instead of resource keys where the agent runs inside Azure.
5. Cache the voice list per region, since it returns hundreds of entries at once.

Rime TTS API + MCP

1. Always send modelId, since a missing one routes to Mist v3 and Coda speakers then fail.
2. Keep HTTP requests under 1,000 characters or use the WebSocket.
3. Back off on 429 before reopening a WebSocket, and remember each retry bills again.
4. Use /ws3 when you need word timestamps.
5. Keep the key server-side, since WebSockets take it in a header and there are no short-lived tokens.