sam-tts

What This Skill Does

The sam-tts skill integrates the legendary Software Automatic Mouth (SAM) synthesizer into the OpenClaw agent ecosystem. SAM is the classic text-to-speech engine famously used on the Commodore 64, providing a distinct, retro-robotic aesthetic that is perfect for projects requiring a nostalgic or synthesized voice tone. The skill provides a robust wrapper around the SAM binary, allowing users to toggle a global 'SAM Mode' where all agent responses are synthesized into WAV audio files, or perform one-off voice commands. It features granular control over vocal characteristics including pitch, speed, mouth resonance, and throat timbre, allowing for deep customization of the robotic output. By maintaining a local state file, the agent remembers your configuration across sessions, ensuring consistent voice profiles.

Installation

To integrate this skill into your agent, run the following command in your terminal: clawhub install openclaw/skills/skills/fourthdensity/sam-tts. Ensure your environment has Node.js 18 or higher installed, as the wrapper scripts rely on modern asynchronous features. After installation, navigate to the skill directory and run npm install to resolve all necessary dependencies for the SAM binary interface.

Use Cases

This skill is ideal for retro-themed AI personas, nostalgic tech demonstrations, or developers building interfaces that require low-bitrate, high-character voice feedback. It is frequently used in gaming environments to mimic retro computer voices or in creative writing tasks where a 'machine' narrator is required. The ability to toggle the mode on and off makes it highly versatile for scenarios where you need to switch between professional text responses and immersive audio-based robotic interaction.

Example Prompts

1. "/sam on"
2. "/sam pitch 120 and speak a greeting to the crew."
3. "/sam Initiate diagnostic sequence, Captain."

Tips & Limitations

• The speed parameter functions inversely: a lower value (e.g., 50) results in a faster speaking rate, while higher values result in a slower, more deliberate cadence.
• Because SAM generates 8-bit style audio, the output is best suited for artistic or diagnostic soundscapes rather than high-fidelity narration.
• Always use the /sam status command if you are unsure if your agent is currently in voice mode, as it avoids accidental audio generation when you prefer quiet text interaction.
• For advanced users, use the --phonetic flag if the standard input produces unexpected pronunciation, allowing for manual correction of the SAM phoneme sequence.