Development Note: This entire platform was developed exclusively using TRAE.AI IDE powered by elite GLM 4.7 model. Learn more about this architecture here.

Fork Note: This project is a specialized fork of ClavixDev/Clavix, reimagined as a modern web-based platform for visual prompt engineering and product planning.

Transform vague ideas into production-ready prompts and PRDs. PromptArch is an AI orchestration platform designed for software architects and Vibe Coders, featuring a plan-first workflow with multi-provider AI support and live canvas rendering.

Developed by Roman | RyzenAdvanced

Gitea Repository: admin/PromptArch
Live Site: rommark.dev/tools/promptarch
Telegram: @VibeCodePrompterSystem

Core Capabilities

Feature	Description
AI Assist	Plan-first workflow: describe a task, get a structured plan, approve, then generate working code with live preview
Prompt Enhancer	Refine vague prompts into surgical instructions using 9 enhancement strategies and 11+ intent patterns
PRD Generator	Convert ideas into structured Product Requirements Documents
Action Plan	Decompose PRDs into actionable development steps and framework recommendations
Google Ads Generator	Generate ad campaigns with XLSX and HTML report export
Slides Generator	Create presentation decks from prompts
Market Researcher	AI-powered market research and analysis

Features

Plan-First Workflow (v1.3.0)

AI generates a structured plan (architecture, tech stack, files, steps) before any code
Plan Review Card with Modify Plan, Start Coding, and Skip to Code actions
After code generation: Preview canvas + Request Modifications buttons
Streaming plan mode with real-time parsing and canvas suppression

Multi-Provider AI (4 Providers)

Provider	Auth	Models
Qwen Code	OAuth (2,000 free req/day)	Qwen Coder models
Ollama Cloud	API Key	Open-source models
Z.AI Plan	API Key	GLM general + coding models
OpenRouter	API Key	20+ models (Gemini, Llama, Mistral, etc.)

Visual Canvas

Live code rendering with [PREVIEW] tags
HTML, React, Python, and more — rendered in-browser
Auto-detect renderable vs. code-only previews
Responsive preview with device size selector (Full / Desktop / Tablet / Mobile)

Code Review & Web Search

Review Code — Send generated code back to AI for bug/security/performance review
Web Search Grounding — Toggle to enrich prompts with live web search results via SearXNG

Enhanced Prompt Engine

9 strategies: clarify, add-context, add-constraints, structure, add-examples, set-tone, expand, simplify, chain-of-thought
Context-aware strategy selection based on detected intent
11+ intent detection patterns (coding, creative, analysis, etc.)

Other

Multi-language support (English, Russian, Hebrew)
Download generated artifacts as ZIP
Push to GitHub integration
Resilient multi-tier provider fallbacks

Quick Start

Clone & Install:

git clone https://github.rommark.dev/admin/PromptArch.git
cd PromptArch
npm install

Configuration: Copy .env.example to .env and add your API keys:
```
cp .env.example .env
```
Configure at least one provider:
- Qwen: Get OAuth credentials from qwen.ai
- Ollama: Get API key from ollama.com/cloud
- Z.AI: Get API key from docs.z.ai
- OpenRouter: Get API key from openrouter.ai/keys (free tier available)
Launch:
```
npm run dev
```
Open http://localhost:3000 to begin.

Tech Stack

Framework: Next.js 15 (App Router, Turbopack)
Styling: Tailwind CSS
State Management: Zustand
Components: shadcn/ui (Radix UI)
Icons: Lucide React
Markdown: react-markdown
Language: TypeScript

Project Structure

promptarch/
  components/
    AIAssist.tsx          # Main AI chat with plan-first workflow (1453 lines)
    PromptEnhancer.tsx     # Prompt enhancement UI with intent detection (556 lines)
    SettingsPanel.tsx      # Provider configuration and API key management (569 lines)
    Sidebar.tsx            # Navigation sidebar
    GoogleAdsGenerator.tsx # Google Ads campaign generator
    PRDGenerator.tsx       # Product Requirements Document generator
    ActionPlanGenerator.tsx # Action plan decomposition
    SlidesGenerator.tsx    # Presentation deck generator
    MarketResearcher.tsx   # Market research tool
    HistoryPanel.tsx       # Chat history management
  lib/
    enhance-engine.ts      # Modular prompt enhancement (9 strategies)
    store.ts               # Zustand state store
    artifact-utils.ts      # Preview/rendering utilities
    export-utils.ts        # Export to XLSX/HTML/ZIP
    services/
      qwen-oauth.ts        # Qwen OAuth streaming service
      ollama-cloud.ts      # Ollama Cloud streaming service
      zai-plan.ts          # Z.AI Plan streaming service
      openrouter.ts        # OpenRouter streaming service
      model-adapter.ts     # Unified provider adapter
      adapter-instance.ts  # Provider registry
    i18n/
      translations.ts      # EN/RU/HE translations
  types/
    index.ts               # TypeScript interfaces

Versioning

This project follows Semantic Versioning. See CHANGELOG.md for detailed release notes.

Version	Date	Highlights
1.7.0	2026-03-18 20:44 UTC	Industry-grade SEO audit, plan flow fix for non-code agents
1.6.0	2026-03-18 20:34	SEO web audit, URL fetching, auto web search for SEO mode
1.5.0	2026-03-18 20:29	Modification progress overlay, preview blink fix
1.4.0	2026-03-18 19:57	Review Code button, web search grounding, responsive preview, model selector fix
1.3.0	2026-03-18 18:51	Plan-first workflow, OpenRouter, post-coding UX, enhanced prompt engine
1.2.0	2026-01-19 19:16	SEO agent fixes, Z.AI API validation
1.1.0	2025-12-29 17:55	GitHub push, XLSX/HTML export, OAuth management
1.0.0	2025-12-29 13:51	Initial release

Development

npm install          # Install dependencies
npm run dev          # Development server (Turbopack)
npm run build        # Production build
npm start            # Start production server
npm run lint         # Lint code

Attribution & Credits

Author: Roman | RyzenAdvanced

Gitea: admin/PromptArch
Telegram: @VibeCodePrompterSystem

Forked from: ClavixDev/Clavix

Visual and architectural evolution of the Clavix framework

Development Platform: TRAE.AI IDE powered by GLM 4.7

License

ISC

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.