sapi-tts

What This Skill Does

The sapi-tts skill is a high-performance, lightweight bridge between OpenClaw AI and the Windows SAPI5 Speech API. Unlike modern cloud-based or GPU-reliant text-to-speech models that require significant VRAM and internet connectivity, this tool leverages the native Windows architecture. It provides near-zero latency, making it ideal for real-time AI responses, system notifications, or accessibility features on Windows 10 and 11 workstations. By accessing locally installed Neural voices, it provides high-quality, natural-sounding audio synthesis with no performance overhead.

Installation

1. Locate your OpenClaw skills directory.
2. Create a new file named tts.ps1 and paste the provided PowerShell script into it.
3. Ensure your execution policy allows local scripts (Set-ExecutionPolicy RemoteSigned -Scope CurrentUser).
4. Verify your installation by running the command: clawhub install openclaw/skills/skills/getmolty/sapi-ttsr.
5. Confirm your voices are detected by running the script with the -ListVoices flag to see your available Neural and Legacy options.

Use Cases

This skill is perfect for developers building autonomous AI agents that need to provide immediate audio feedback. Use it to build an 'AI Voice Assistant' that speaks answers instead of just typing them. It is also excellent for batch processing text files into spoken audio for accessibility, or for system administrators who want to trigger audio alerts based on log monitoring. Because it runs locally, it is ideal for environments where privacy is paramount and data should not be sent to cloud TTS providers.

Example Prompts

1. "Speak the following text: 'System update completed successfully' using my default voice."
2. "Read the summary of this project file aloud to me."
3. "List all my available SAPI5 voices and tell me which one is best for English."

Tips & Limitations

To get the best quality, ensure you have the 'Microsoft Natural Voices' packs installed in your Windows Settings under 'Speech'. The quality of the output is strictly dependent on the voice engine installed in your OS; standard legacy voices may sound robotic compared to modern neural ones. The script allows for real-time speed adjustments using the -Rate parameter, which is useful for audio-skimming tasks. Note that this tool is Windows-exclusive; it will not function on Linux or macOS environments.