@ Designed for people who don’t ask AI questions — they architect intelligence , @LegendWorks , @Stay Updated
Modern AI systems are only as powerful as their prompts. AI Architect transforms vague ideas into production-ready prompt blueprints — ideal for:
Developers
Researchers
Startup founders
Prompt engineers
AI product teams
This is a Prompt Engineering IDE.
| Feature | Description | Status |
|---|---|---|
| Persona-Driven Prompting | Generate prompts using professional AI personas | ✅ Stable |
| CO-STAR Architecture | Context, Objective, Style, Tone, Audience, Response | ✅ Implemented |
| Chain-of-Thought Control | Step-by-step reasoning enforcement | ✅ Optional |
| Few-Shot Examples | Example-driven outputs | ✅ Optional |
| Iterative Self-Correction | Output validation before final response | ✅ Optional |
| Negative Constraints | Explicit exclusions to prevent hallucinations | ✅ Implemented |
| XML-Structured Prompts | High-precision prompt boundaries | ✅ Implemented |
| Offline Execution | No API, no internet dependency | ✅ Yes |
| Persona | Best Used For |
|---|---|
| Software Engineer | System design, APIs, backend logic |
| Python Expert | Optimization, clean code, libraries |
| Senior Data Scientist | ML pipelines, statistics, research |
| Product Manager | Roadmaps, PRDs, strategy |
| CEO / Business Strategist | Decision-making, market analysis |
| UX/UI Designer | User flows, usability, accessibility |
| Cybersecurity Analyst | Threat modeling, secure systems |
| Academic Researcher | Papers, citations, structured analysis |
| Parameter | Available Options |
|---|---|
| Tone | Professional, Technical, Academic, Casual, Persuasive |
| Complexity | ELI5, Standard, Advanced, Expert |
| Output Format | Markdown, Bullets, Code, JSON, CSV, Tables |
| Reasoning Model | Chain-of-Thought, Few-Shot, Iterative |
| Logic Model | Description | Benefit |
|---|---|---|
| Chain-of-Thought | Forces step-by-step reasoning | Reduces logical errors |
| Few-Shot Examples | Demonstrates expected output | Improves alignment |
| Iterative Review | Self-checks output quality | Minimizes hallucinations |
| Feature | AI Architect | Typical Prompt Tool |
|---|---|---|
| Offline | ✅ Yes | ❌ No |
| Prompt Structure | XML + CO-STAR | Free-text |
| Reasoning Control | Advanced | Limited |
| Persona Control | Yes | Partial |
| Professional UI | Desktop-grade | Web forms |
| User Type | Use Case |
|---|---|
| Developer | Generate deterministic coding prompts |
| Researcher | Structured academic queries |
| Startup Founder | Business analysis & planning |
| Prompt Engineer | Prompt blueprint creation |
| Educator | Clear instructional prompts |
Software Engineer
Data Scientist
CEO
Legal Expert UX Designer
Cybersecurity Analyst & more
Chain-of-Thought
Few-Shot Examples
Iterative Self-Correction
Tone
Complexity
Output Format
XML-tagged sections for LLM reliability
Modern Tkinter UI (Dark output console style)
Copy-ready prompt blueprints
⚙️ Offline ,No API keys, no internet required
SYSTEM IDENTITY │ ├── CONTEXT │ └── Background & intent │ ├── OBJECTIVE │ └── Core task definition │ ├── CONSTRAINTS & INSTRUCTIONS │ ├── Rules │ ├── Reasoning model │ ├── Negative constraints │ └── OUTPUT SPECIFICATIONS ├── Tone ├── Format └── Complexity
This structure significantly reduces hallucination and improves determinism across LLMs.
Modern control panel + terminal-style output
Left: Prompt configuration & logic controls
Center: Project specification inputs
Right: Generated prompt blueprint
(Add screenshots here later)
Requirements Dependency Version Python 3.9+ Tkinter Bundled with Python OS Windows / Linux / macOS Run Locally git clone https://github.com/LegedsDaD/AI-Architect.git cd AI-Architect python prompt_generator.py
No virtual environment required.
###SYSTEM IDENTITY You are acting as a Senior Data Scientist. Your response complexity must be Advanced.
###CONTEXT You are designing an ML pipeline for healthcare data.
###OBJECTIVE Design a scalable feature engineering workflow.
###BEGIN RESPONSE
Language - Python UI Framework - Tkinter (ttk themed) Design Pattern - MVC-like separation Prompt Design - CO-STAR + XML tagging
We welcome serious contributions.
Fork the repository
Create a feature branch
Follow clean-code practices
Submit a Pull Request
Bonus points for UX improvements, logic engines, or prompt science research.
This project is licensed under the MIT License — free to use, modify, and distribute.
If this tool helped you think better, design better, or prompt smarter — give it a ⭐ and share it with other developers.
LegedsDaD
#Designed for people who don’t ask AI questions — they architect intelligence.