-
-
The scene illustrates AskOne’s mission to empower individuals with a visually structured and intuitive interface.
-
AskOne – representing the system’s role as a decision-making co-pilot for creators, researchers, and innovators."
-
AskOne’s AI Core – The iconic ‘A’ rendered as a living circuit, symbolizing intelligence, ethics, and adaptive computation
-
This showcases the system's clean interface, highlighting modules like context analysis, ethical guardrails, and domain classification.
Inspiration
AskOne was built in under 12 hours — but the spark that created it was forged over a decade of digital curiosity, search, and self-evolution.
This project began as more than just a hackathon submission — it was a convergence of everything I’ve searched for, studied, and dreamed of using my personal Google account over the past 10 years. From philosophical inquiries to technical rabbit holes, every search became a thread that led to this moment — a universal AI agent capable of turning a single thought into structured possibility.
I used Gemini 2.5 Flash as a catalyst — a model so powerful it could help me synthesize that lifetime of knowledge and channel it into a single architectural vision. Together with OpenAI’s ChatGPT, I built the core logic, refined the system prompts, and layered meaning, utility, and precision into every part of AskOne.
This wasn’t just co-building with AI — it was co-evolving with it.
AskOne is the result of: • 🔍 Years of human curiosity • ⚡ Hours of inspired execution • 🧠 Real-time collaboration between human intuition and machine intelligence
It’s a tribute to the idea that sometimes, with the right tools, one prompt really is enough.
What it does
AskOne is a universal AI agent that transforms a single prompt into structured, actionable, and ethically sound solutions. Powered by Gemini 2.5 Pro, AskOne identifies the user’s intent, classifies the domain (e.g., business, personal, creative), and delivers tailored responses using adaptive frameworks like Strategist, Planner, Coach, Engineer, or Creator. Whether you’re facing a business dilemma, a technical challenge, or a creative block, AskOne provides guidance, solutions, and next steps — all from one prompt.
How we built it
We built AskOne on Replit using Node.js, Express, and the Gemini 2.5 Pro API. The backend parses user prompts and returns structured responses using a custom system prompt designed for ethical reasoning and domain-aware replies. The frontend is fully mobile-optimized and includes an interactive prompt box, category selector, and output area. We integrated environment variables to securely handle the API key and ensured smooth UI/UX with custom styling and animations.
Challenges we ran into
Finding the balance between flexibility and clarity in responses from the Gemini API • Ensuring the agent could interpret vague or complex prompts without user frustration • Optimizing the mobile experience while maintaining a futuristic, judge-ready aesthetic • Managing fast iteration speed while building with layered UI/UX and multi-agent logic • Deciding between stability (Gemini 1.5 Pro) and innovation (Gemini 2.5 Pro) for final deployment
Accomplishments that we're proud of
Built AskOne from concept to working MVP in less than 12 hours • Integrated Gemini 2.5 Pro with advanced prompt control and real-time response routing • Designed a futuristic UI with full mobile compatibility and intuitive interaction • Created a modular architecture that supports multiple agent styles (Strategist, Planner, Creator…) • Launched AskOne under a secured domain (AskOne.xyz) and published a live demo
What we learned
The power of system prompts and controlled prompting with Gemini models • How to design AI assistants with ethical safeguards and contextual awareness • How small UX touches (like tap feedback and layout clarity) massively impact perception • That building under pressure often produces our best ideas • And most importantly: sometimes “One Prompt” really is enough to change everything
What's next for AskOne - Universal AI Agent for Cross-Domain Problem Solving
Add persistent memory using Firebase or a lightweight vector database • Expand the multi-agent framework with deeper specialization (e.g. Legal, Wellness, Scientific) • Train AskOne on example user journeys to improve contextual clarity • Launch a public beta and explore monetization (AskOne Pro Tier) • Explore integrating AskOne into platforms like Notion, Google Docs, or Discord for real-time collaboration • Use AskOne as the foundation for a full OS layer for everyday decision-making and AI-augmented thought
Log in or sign up for Devpost to join the conversation.