createOpenAISpeech (src/speech.ts)

* speech.ts — text-to-speech (2.0). `generateSpeech` is the audio sibling of `generateImage`: one request, one answer, no streaming and no canonical delta stream — a TTS endpoint returns a finished audio file, not a token sequence. Two wires ship: OpenAI's `POST /audio/speech` an

// Source: https://github.com/Deuz-AI/Deuz-SDK | License: MIT | Copyright 2026 Umutcan Edizaslan (U-C4N / Deuz)
// Provided via justclaw.it.com; keep this notice with the code.
/**
 * speech.ts — text-to-speech (2.0).
 *
 * `generateSpeech` is the audio sibling of `generateImage`: one request, one
 * answer, no streaming and no canonical delta stream — a TTS endpoint returns
 * a finished audio file, not a token sequence. Two wires ship: OpenAI's
 * `POST /audio/speech` and ElevenLabs' `POST /text-to-speech/{voice}`; both sit
 * behind the `SpeechAdapter` seam in `adapters/speech.ts`.
 *
 * PURE + edge-safe: HTTP goes through the injected `deps.fetch` and the API key
 * through `deps.keyProvider` / factory config / `ClientConfig.apiKeys` (the G1
 * precedence chain) — never read from the e

... (truncated -- full source via MCP)