Speak

What This Skill Does

The Speak skill provides an adaptive layer for Text-to-Speech (TTS) interactions within the OpenClaw environment. Rather than relying on static, generic voice profiles, this skill employs a feedback-loop mechanism that observes user preferences regarding tone, speed, verbosity, and voice provider selection. By monitoring how a user reacts to previous outputs, the agent autonomously tunes its delivery style to better match the user's personality or communication constraints. It maintains a stateful configuration that captures specific voice providers, stylistic traits, and negative constraints to ensure every interaction feels personalized and intuitive.

Installation

To integrate the Speak skill into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/ivangdavila/speak

Ensure that you have completed the base TTS setup as detailed in your local config.md before installation. The skill will begin observing interactions immediately upon activation.

Use Cases

• Accessibility Optimization: Automatically adjusting pitch or cadence for users who find standard TTS difficult to parse.
• Professional vs. Casual Toggling: Switching from concise, information-dense reports to conversational, empathetic responses based on user context.
• Preference Mirroring: Detecting if a user prefers direct command-based communication or flowery, conversational explanations, and adapting the agent's persona to match.
• Constraint Enforcement: Dynamically updating the 'Avoid' list to prevent the agent from using specific jargon or intrusive vocal patterns.

Example Prompts

1. "Speak, I find that voice a bit too robotic; please switch to a warmer, more natural tone for future updates."
2. "From now on, keep your spoken responses extremely brief; just give me the facts and skip the introductory pleasantries."
3. "Stop using that high-pitched voice profile; stick to the bass-heavy synthesizer we used last Tuesday."

Tips & Limitations

• Patience for Evolution: The skill requires at least two consistent signals to confirm a preference change. Do not expect immediate, permanent configuration shifts after a single request.
• Manual Overrides: Always check the criteria.md file if the auto-evolution behaves unexpectedly. You can manually edit the generated settings if you need to force a specific configuration.
• System Compatibility: Ensure your host operating system has the necessary drivers for the TTS providers defined in your configuration, as the skill only manages the logic of selection and stylistic mirroring.