Introduction
Welcome to Prism Context Engine — The Context Operating System for Vibecoders.
New to Prism? Start with Quick Start to get up and running in 5 minutes.
What is Prism Context Engine?
Prism Context Engine is an AI-powered context governance platform that eliminates “Context Pollution” — the phenomenon where AI coding assistants generate code that doesn’t match your project’s architecture, style, or patterns.
The Problem
Every AI coding assistant (Cursor, Windsurf, Copilot, Claude) has the same issue:
You: "Create a login form"
AI: *generates Bootstrap-looking component with inline styles*
You: "No, use our design system with glass morphism and cyan accents"
AI: *still doesn't quite get it*Context Pollution happens when AI doesn’t understand:
- Your design system rules
- Your architectural patterns
- Your naming conventions
- Your error handling strategies
- Your testing preferences
The Solution
Prism Context Engine captures your architectural decisions from video recordings and converts them into actionable context rules that AI assistants can follow.
How It Works
Step 1: Record Your Architecture
Screen record yourself explaining your project’s patterns, design system, and conventions. Just talk naturally — Prism handles the rest.
Step 2: AI Extracts Rules
Azure OpenAI analyzes your video transcript and automatically extracts:
- Coding patterns and conventions
- Design system rules
- Component architecture decisions
- Testing strategies
Step 3: Deploy to Your IDE
Export rules as .cursorrules, .windsurfrules, or connect via MCP (Model Context Protocol) for real-time sync.
Step 4: AI Follows Your Rules
Your AI assistant now understands your architecture. No more generic Bootstrap code.
The Core Metaphor
| Concept | Prism Equivalent | Description |
|---|---|---|
| 🍳 The Kitchen | SaaS Dashboard | Where you (the Chef) mix ingredients and define recipes |
| 📜 The Recipe | Context Rules | The strict instructions your AI must follow |
| 🍽️ The Waiter | MCP Protocol | Delivers fresh recipes directly to your IDE |
Key Features
Video-to-Context
Upload screen recordings, get structured rules automatically
MCP Integration
Real-time sync with Cursor, Windsurf, VS Code, and Claude
Semantic Search
Find relevant architectural discussions using AI-powered search
Brand Profiles
Maintain consistent design systems across projects
Supported Editors
| Editor | Status | Integration Method |
|---|---|---|
| Cursor | ✅ Full Support | .cursorrules + MCP |
| Windsurf | ✅ Full Support | .windsurfrules + MCP |
| VS Code | ✅ Full Support | MCP Extension |
| Claude Desktop | ✅ Full Support | MCP |
| GitHub Copilot | 🚧 Coming Soon | — |
Architecture Overview
┌─────────────────────────────────────────────────────────────┐
│ Prism Dashboard (Web) │
│ • Video Upload (Mux) │
│ • Rule Management │
│ • Brand Profiles │
└────────────────────────────┬────────────────────────────────┘
│ API / Webhook
▼
┌─────────────────────────────────────────────────────────────┐
│ Azure OpenAI (GPT-4o-mini) │
│ • Transcript Analysis │
│ • Rule Extraction │
│ • Embeddings for Semantic Search │
└────────────────────────────┬────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────┐
│ Azure Cosmos DB (MongoDB API) │
│ • Rules Collection │
│ • Video Transcripts │
│ • Vector Embeddings │
└────────────────────────────┬────────────────────────────────┘
│ MCP Protocol
▼
┌─────────────────────────────────────────────────────────────┐
│ Your IDE (Cursor/Windsurf/VS Code) │
│ AI assistant automatically follows your rules │
└─────────────────────────────────────────────────────────────┘Next Steps
Installation
Set up your account and connect your IDE
Quick Start
Create your first context rule in 5 minutes
Current Version: 1.0.3
Built by: JeffDev Studio