Inspiration
What it does
How we built it
Challenges we ran into
Accomplishments that we're proud of
Inspiration
In an era where AI can synthesize any voice in seconds, we questioned: who
truly owns your acoustic identity? As we move toward an AI-agent economy,
voices are becoming our most valuable digital signatures. We were
inspired to build a platform that doesn't just "process" voice, but
empowers it—giving creators and global citizens a way to transform their
vocal expression while maintaining sovereign ownership through
decentralized infrastructure.
What it does
VOISSS Butler is a Flutter-powered digital assistant that allows users to
record, transform, and secure their voices on-chain. It features:
* 🎭 AI Voice Transformation: Morph your voice into high-fidelity AI
personas using advanced voice synthesis
* 🎙️ Intelligent Butler Assistant: A context-aware AI companion powered by
Gemini 3.0 Flash that provides insights and manages recordings through
agentic voice commands
* ⛓️ Decentralized Storage: Every recording is stored as a secure, on-chain
artifact on IPFS and indexed via blockchain networks
* 🧠 AI Insights: Multi-modal analysis of recordings to generate
transcripts, summaries, and tags automatically
* 📝 Transcript Composer: A tool to create shareable video transcripts
directly from your AI-transformed audio
How we built it
The stack is a cutting-edge fusion of frontier AI and Web3 infrastructure:
* Backend: Serverpod 3 for scalable Dart-native backend services
* Frontend: Flutter for cross-platform mobile experiences
* AI Intelligence: Integrated Google Gemini 3.0 Flash for lightning-fast
reasoning, context-aware assistance, and multimodal insights
* Vocal Synthesis: Leveraged ElevenLabs for professional-grade TTS and
voice cloning to ensure emotional resonance and native accents
* Web3 Backbone: Built on blockchain infrastructure for gasless, high-speed
on-chain recording provenance
* State Management: Provider and BLoC patterns for reactive state
management
Challenges we ran into
Integrating real-time voice interaction with decentralized protocols
presented significant latency hurdles. Ensuring the AI Butler could
respond naturally while simultaneously handling blockchain state and IPFS
uploads required a robust, asynchronous architecture. Coordinating the
"Agentic" capabilities of Gemini—allowing it to trigger UI actions across
the app—required a deep dive into state synchronization and global context
management. Additionally, building a seamless Serverpod backend that
could handle the complex audio processing pipeline while maintaining
low-latency responses was challenging.
Accomplishments that we're proud of
We are incredibly proud of achieving Gemini 3.0 "Agentic" integration,
where the assistant doesn't just talk, but can actually control the
application's navigation and workflow based on spoken intent. Successfully
implementing a gasless on-chain storage flow that feels as smooth as a
standard web app is another major win, bridging the gap between Web2 UX
and Web3 security. Finally, achieving a production-ready Flutter interface
with real-time audio processing capabilities within the hackathon
timeframe was a significant milestone.
What we learned
We learned that the true power of AI agents lies in Context Awareness. By
feeding the assistant real-time page data and recording metadata, it
transformed from a simple chatbot into a genuine productivity partner. We
also deepened our understanding of the L2 ecosystem, particularly how to
optimize metadata for on-chain permanence while maintaining the speed
users expect from modern voice applications. Working with Serverpod taught
us the benefits of having a unified Dart ecosystem for both frontend and
backend development.
What's next for VOISSS
The roadmap includes Autonomous Dubbing Pipelines where the AI can
translate and dub entire video content in one click. We are also pursuing
deep Farcaster Integration to allow "Voice Moments" to be shared directly
across the decentralized social graph. Lastly, we plan to release the
VOISSS Creator SDK, allowing other developers to build voice-powered,
decentralized applications on top of our Serverpod-powered protocol.
Log in or sign up for Devpost to join the conversation.