Synthesizes speech audio from text using the specified provider and voice.

POST 
/studio/speech

Records an OrcaRequest, estimates token usage/cost, persists the generated audio, and returns media metadata.

Processing steps:

1. Validate inputs (text, provider).
2. Resolve the provider adapter from configured speech groups.
3. Create and persist an OrcaRequest in InProgress state.
4. Estimate prompt/completion tokens and cost via provider-specific estimators; save to request.
5. Invoke synthesis (SynthesizeSpeechAsync), receive base64 audio.
6. Persist audio as OrcaAssetType.Speech and return its metadata.

Always responds with HTTP 200 for handled outcomes; clients should inspect the Error field.

Request

Responses

200

OK

401

Unauthorized - missing or invalid credentials.

403

Forbidden - authenticated but missing required role or policy.