Skip to content
TheChatArchiveThe Chat Archive
MiniMaxminimax-m2.7Hugging Face Arena WebDev Preferences (2026)

Crea una app que transforme un prompt del usuario, en un prompt ajustado para ser leído…

  1. User

    Crea una app que transforme un prompt del usuario, en un prompt ajustado para ser leído por inteligencia artificial. El usuario debe poder elegir el modelo objetivo para usar el prompt, y las respuestas deben elaborarse en base a los últimos métodos de prompt engineering, tanto oficiales como comunitarios. La [SECURITY_DATA_1] de modelos debe tener al menos los 5 modelos más recientes de openai, anthropic, grok, deepseek y kimi. Debe tener función por dictado de voz, para que el usuario explique con lujo y detalle. Debe detectar inglés y español. Asegurate de que tenga al menos optimización para fable, gemini 3.6 y gpt 5.6. Give it a beautiful and colorful pixel-art style

  2. MiniMax Assistant

    I'll create a stunning pixel-art prompt engineering app. Let me start by creating the specification document.

    Now let me create the main application with all the features:

    Let me fix the TypeScript errors by adding proper type declarations:

    Let me fix the TypeScript issues more comprehensively:

    Now let me update the index.html with the correct title and fonts:

    Now let me update the Tailwind CSS config to add the custom fonts and animations:

    Now let me build the project to verify everything works:

    The project builds successfully! Let me verify the final implementation by reading the complete App.tsx:

    The app is complete and builds successfully! Let me provide a summary of what was created:

    ✅ Prompt Architect - AI Prompt Engineering Studio

    Features Implemented:

    🤖 AI Model Selection:

    • OpenAI: GPT-4o, GPT-4o Mini, GPT-4 Turbo, o1 Preview, o1 Mini, GPT 5.6
    • Anthropic: Claude 3.5 Sonnet, Claude 3.5 Haiku, Claude 3 Opus, Claude 3 Sonnet, Claude 3 Haiku
    • Grok: Grok 3 Beta, Grok-2, Grok-1.5, Grok-2 Mini, Grok-1
    • DeepSeek: DeepSeek V3, DeepSeek Coder V2, DeepSeek Math, DeepSeek VL2, DeepSeek Prover
    • Kimi: Kimi K2, Kimi K1.5, Kimi K1, Kimi Moonshot, Kimi Vision
    • Google: Gemini 3.6, Gemini 2.0 Flash, Gemini 1.5 Pro, Gemini 1.5 Flash
    • Meta: Fable, Llama 3.3 70B, Llama 3.2

    ⭐ Special Optimizations:

    • Fable: Interactive storytelling with scene markers, choice systems, character consistency
    • Gemini 3.6: Thinking budget protocol, tool integration, code execution patterns
    • GPT 5.6: Agentic workflow framework, memory utilization, advanced reasoning

    🎤 Voice Dictation:

    • Real-time speech recognition using Web Speech API
    • Auto language detection (English/Spanish)
    • Continuous mode for detailed explanations

    📚 Prompt Engineering Techniques (10 total):

    • Chain of Thought (CoT)
    • Zero-shot / Few-shot Learning
    • Role Prompting
    • Constitutional AI
    • Tree of Thoughts (ToT)
    • Output Formatting
    • System Context Enhancement
    • Parameter Hints (Temperature)
    • DSPy Methods

    🎨 Pixel-Art Design:

    • Retro 8-bit aesthetic with neon colors
    • Animated starfield background
    • CRT scanline effect
    • Custom pixel fonts (Press Start 2P, VT323, Fira Code)
    • Glowing effects and pulse animations
    • Dark space theme with cyber accents

    🌐 Bilingual Support:

    • Full English and Spanish translations
    • Auto language detection
    • Localized prompts and UI

    Files built: SPEC.md, src/App.tsx, index.html, src/index.css