Summary

The TTS guide does not show how to use the existing OpenAI-compatible and local CLI providers with fully local speech engines on macOS, Linux, or Windows.

Affected docs path(s) or URL(s)

docs/tools/tts.md
https://docs.openclaw.ai/tools/tts

Steps to reproduce / verify

1. Open the TTS guide.
2. Review the OpenAI-compatible endpoint and Local CLI configuration sections.
3. Search for Speech Swift, speech-core, macOS, Linux, or Windows local speech setup.
4. Observe that no complete local configuration recipe is provided.

Expected docs behavior/content

The guide should provide copyable configurations for Speech Swift over its loopback OpenAI speech endpoint on macOS and for the existing Local CLI provider on macOS, Linux, and Windows. It should also document the explicit OpenAI responseFormat override needed by WAV-only local endpoints.

Actual docs behavior/content

The guide documents generic custom OpenAI endpoints and Local CLI placeholders, but users must independently discover the correct executable, arguments, WAV format, model setup, and platform-specific runtime.

Impact

Users who want private, offline TTS have the required OpenClaw provider surfaces already, but cannot configure them from the published guide. This adds avoidable trial and error and makes the existing cross-platform local path difficult to discover.

Evidence

• Speech Swift OpenAI-compatible speech endpoint: feat: add OpenAI-compatible speech synthesis soniqo/speech-swift#415
• speech-core Linux CLI reference: https://github.com/soniqo/speech-core/blob/main/docs/cli.md
• speech-core Windows setup: https://soniqo.audio/getting-started/windows
• Existing OpenClaw configuration surface: docs/tools/tts.md