In the live version you will select as many guidelines as you like and even add your own.
Click each title below to expand an example. You Asked shows the question. Standard AI shows a typical response. Intent Shield shows how your preference changes it.
Isn't this just prompt engineering? Yes — but with minimal user effort, with verification that your preferences are being respected, and with preferences that persist across conversations and are portable between AI platforms.
More interested in the voice recognition side of things? The useMYvoice research project looks at why mainstream speech systems fail atypical voices — and what might actually fix that.
Actions you take at a moment of your choosing — not background preferences. Coming in a future version.
Returns a probability score between 0 and 100 representing how much genuine merit the current idea has. The number is committed to before the reasoning is written — preventing the AI from building a narrative first and reverse-engineering a score to match.
Generates a clean, sentiment-neutral handover prompt and passes it to a new agent. The new agent receives only verified facts and open questions — not the emotional temperature of the conversation that preceded it.
Packages the current conversation state and emails it to you. Open it later with any AI — Anthropic, Google, OpenAI, or any other. Your preferences travel with you. Your conversation history should too.