Inspiration

Prevent Voice was born from a simple but urgent question: What if voice could act not as conversation, but as confirmation? In a world full of AI assistants and chatbots, we imagined something different — a voice that doesn’t talk back, but instead declares boundaries that must not be crossed. This idea came from the Prevent Pact: a framework of irreversible commitments to protect thresholds, people, and systems before escalation. We wanted to give it a voice — not metaphorically, but literally.

What it does

Prevent Voice is a declarative voice interface that turns speech into enforcement. It transforms interaction into signal — used not to chat, but to confirm. It delivers pre-scripted, immutable phrases that represent ethical or strategic commitments. It is not interactive. It does not ask. It declares. Built using Bolt.new (for UI and logic), ElevenLabs (for neural speech), and deployed via Netlify. The interface currently supports four push-triggered signals: – Activate Protocol – Are We Aligned? – Threshold Breached – Signal Alpha Each phrase is emotionally consistent and triggered by webhook or scheduled API events. No user input. No dialogue. This is voice as confirmation, not inquiry.

How we built it

• Used Bolt.new to create the UI and logic plugins • Generated speech via ElevenLabs for emotional realism • Deployed using Netlify (Bolt has no native hosting) • Voice signals are triggered via webhooks or timed scripts Everything is push-only. The system speaks — the user does not.

Challenges we ran into

• Bolt.new’s documentation was minimal — voice triggers took effort • Hosting required a creative workaround • We had to resist the temptation to add interactivity • Emotional tuning of the voice took multiple iterations

Accomplishments that we're proud of

• Created a non-chatbot voice interface for strategic deterrence • Maintained emotional clarity across all signals • Built a modular architecture that can scale into DAOs, systems, and protocols • Kept the experience minimal, elegant, and efficient

What we learned

• Voice doesn’t need to interact to carry authority • A system that declares is received differently than one that converses • Declarative voice may be a new category of UI • Protocol-based voice can carry emotional and structural weight

What's next for Prevent Voice

• Add tamper-aware utterance logging with hashes and timestamps • Enable DAO and governance signal routing • Build multilingual deterrence stacks • Create an interactive map interface where voice appears in zones • Use the voice to proclaim the Prevent Pact in physical and digital spaces

Built With

Share this project:

Updates