“As everything becomes more complex, we believe in the power of simplicity.”

Foreword
SponeAI is a simple, fast, and central artificial intelligence experience developed with the concept of “single page intelligence”. It offers the user different artificial intelligence capabilities in one screen with an intuitive flow instead of dozens of menus and transitions.

Its vision is ambitious but not flashy: speed, accuracy, cost awareness and accessibility. SponeAI makes access easier instead of competing with large systems, offering an intelligence tool that everyone can use.
Overview
• React, TypeScript and Vite core, Tailwind UI provides a simple interface.
• Firebase provides identity, database, storage and serverless functions.
• Google Gemini (2.0, 2.5 Flash, TTS) and OpenAI GPT-5 run behind a proxy, keys are not leaked to the client.
• Web, Android and iOS share the same codebase with Capacitor.
• Membership and billing: Web (Stripe Billing), Android (Google Play Billing), iOS (App Store In-App Purchases). Same tiers and permissions on all platforms.
Design
Single screen, clear hierarchy and low cognitive load. SponeAI avoids modal mazes and deep navigation. Core actions are in the Action Bar, content flows and is editable in the same place. Typography, contrast and spacing follow an “accessibility first” approach.

• Single page flow, minimal context switching and instant feedback.
• Streaming responses, sanitized HTML and robust code rendering.
• Responsive design, keyboard friendly controls and reduced motion on mobile.
Action Bar, Modals
Quiz
Generates dynamic questions with Gemini 2.5 Flash. Tracks correctness, time and score in Firestore. You can challenge friends and compare results.
Transcript
Transcribes up to 20 MB of audio to text with speaker cues and language codes. Monthly minute quotas are tracked in the transcriptUsage collection and can be increased with add-ons.
Podcast
Generates natural monologue or dialogue from text. Generates WAV with gemini-2.5-flash-preview-tts, uploads to Firebase Storage.
Document
Processes PDF, DOCX, TXT and MD. Offers summary, chapter-based analysis and page-based usage control.
Image
Generates 1024×1024 images with Imagen 4.0 Fast. Nano Banana. Gpt Image mini. The credit system balances free and Pro usage.
Prototype
You can write websites, web applications, and web games, and preview them instantly.
Chat Architecture
• /geminiChat and /openaiChat proxy endpoints support real-time SSE streaming.
• Messages are stored in Firestore chats, with real-time listening and client-side sorting.
Source Engine 2.0
• World Bank (inflation, unemployment, GDP, per capita income).
• Country Card (RestCountries – population, capital, flag, region).
• Wikipedia (summaries of people, concepts, events, institutions).
• OpenLibrary (book, author, publisher and print information).
• TMDB (film, series, actor, director and viewing platforms).
• NASA (APOD – astronomy picture of the day).
• YouTube (tutorial / narration videos only in tutorial queries).
• Exchange Rate Card (foreign exchange and gold rates – USD, EUR, GBP, GOLD).
• Weather (city-based today / tomorrow weather report).
• OpenAlex (scientific articles, meta-analyses and evidence-based content).
• StackOverflow (top-voted solutions for code and error questions).
• Crypto Rates (instant price and conversions for BTC, ETH, etc.).
Modes
Fast Mode: Fast, short, and direct; web search off, low token limit.
Deep Think: Analytical and multi-step; streaming on, web search if needed; optimized with SystemPrompts.

As LLMs evolve, SponeAI integrates them in a balanced way, keeps the interface simple, and makes complex technology feel natural.

“sponeai.com, the simplicity of artificial intelligence on a single screen.”