Fabrice

Posted on Jun 19

I Built an Open-Source Prompt Library for Developers, Creators, and AI Power Users

#webdev #ai #productivity #opensource

AI is everywhere now.

Developers use it to debug code. Designers use it to generate concepts. Founders use it to validate ideas. Marketers use it to write campaigns. Students use it to learn faster. Creators use it to produce images, videos, scripts, and music.

But there is still one annoying problem:

Most people are rewriting prompts from scratch every single day.**

The same debugging prompt.

The same landing page prompt.

The same SEO prompt.

The same product strategy prompt.

The same image generation prompt.

The same resume prompt.

The same study prompt.

So I built Prompt Empir — a professional, open-source AI prompt library designed to make high-quality prompts easier to discover, copy, improve, and share.

GitHub repo: https://github.com/gavuvapro/prompt-empire

What is Prompt Empir?

Prompt Empir is a curated open-source collection of reusable prompts for major AI tools and real professional workflows.

It is built for:

Developers
Designers
Founders
Marketers
Students
Researchers
Writers
Content creators
Freelancers
Businesses
AI enthusiasts

The goal is simple:

Make prompt engineering practical, organized, reusable, and open-source.

Instead of keeping your best prompts scattered across Notion, Google Docs, random chats, and browser bookmarks, Prompt Empir gives them a clean home.

Why I built it

There are many AI prompt lists online, but most of them have the same problems:

They are not organized for serious work.
They are hard to search.
They mix beginner prompts with professional prompts.
They are not designed for contribution.
They do not cover enough AI platforms.
They do not feel like real open-source projects.

I wanted something that feels closer to:

Awesome Lists
Hugging Face documentation
Vercel-style design
Tailwind UI quality
GitHub-readable Markdown

A prompt library should not just be a list of random phrases.

A good prompt is a reusable workflow.

It should include context, constraints, role, expected output, use cases, and example results.

That is what Prompt Empir is trying to standardize.

Supported AI platforms

Prompt Empir is not only for ChatGPT.

It includes prompts for many major AI systems, including:

General AI

ChatGPT
Claude
Gemini
Grok
DeepSeek
Perplexity

Coding AI

Cursor
GitHub Copilot
Windsurf
Replit AI
Bolt
Lovable

Image AI

Midjourney
DALL-E
Flux
Stable Diffusion
Leonardo AI

Video AI

Veo
Runway
Kling

Audio AI

ElevenLabs
Suno
Udio

This makes the project useful for people working across text, code, design, image generation, video generation, and audio generation.

Categories inside the library

Prompt Empir is organized by professional use case, not just by tool.

Current categories include:

Development
Frontend
Backend
React
Next.js
Expo
Python
Node.js
APIs
Databases
DevOps
Security
AI Engineering
Design
UI Design
UX Design
Branding
Logo Design
Business
Startups
SaaS
Product Strategy
Marketing
Sales
Writing
Blogging
SEO
Copywriting
Education
Research
Productivity
Psychology
Philosophy
Sports
Finance
Career
Social Media
E-commerce
Creative Writing
Image AI
Video AI
Audio AI

The idea is to make it easy to find prompts based on what you are trying to do.

Not just:

“Give me a ChatGPT prompt.”

But:

“Give me a prompt to review a Next.js API route for security issues.”

Or:

“Give me a prompt to create a SaaS pricing strategy.”

Or:

“Give me a prompt to generate a cinematic Runway video ad shot list.”

Every prompt follows a clean format

Each prompt is written in Markdown and follows a consistent structure:

# Prompt Title

## Description
Explain what the prompt does.

## Best AI Models
- ChatGPT
- Claude
- Gemini

## Use Cases
- Example 1
- Example 2

## Prompt
[prompt here]

## Example Output
[sample output]

This makes the repository easy to read on GitHub and easy to convert into a website.

It also makes contributions easier because everyone follows the same standard.

Example prompt

Here is the kind of prompt structure the project encourages:

# Production API Design Review

## Description
Reviews REST or GraphQL API designs for correctness, security, scalability, and developer experience.

## Best AI Models
- ChatGPT
- Claude
- DeepSeek
- GitHub Copilot

## Use Cases
- Validate endpoint design
- Improve API documentation

## Prompt

You are a principal backend architect. Review this API design for resource modeling, authentication, authorization, validation, pagination, errors, idempotency, rate limiting, observability, and versioning.

Provide a concise scorecard and improved endpoint examples.

API draft:
{{api_draft}}

Business requirements:
{{requirements}}

## Example Output

Scorecard: Authentication 8/10, Authorization 5/10.
Recommendation: add object-level authorization checks for GET /projects/{id} and use cursor pagination for list endpoints.

Notice that this is not just “write an API prompt.”

It gives the AI a role, a task, evaluation areas, context placeholders, and a clear output expectation.

That is what makes prompts more reusable.

The website experience

Prompt Empir is not only a GitHub folder.

It also includes a website built with:

Next.js
TypeScript
Tailwind CSS
Markdown content
Fuse.js search
Lucide icons
Vercel deployment support

The website supports:

Browse by category
Full-text search
Filter by AI model
Copy prompts with one click
Prompt detail pages
Markdown rendering
Dark mode, light mode, and system theme
SEO metadata
Sitemap
Robots.txt
Open Graph previews

This means the project can work both as:

A GitHub open-source repository
A polished public prompt library website

Why open-source?

Prompt quality improves when people can inspect, discuss, edit, and contribute.

Open-source makes this possible.

Anyone can:

Fork the repository
Add prompts
Improve existing prompts
Suggest categories
Fix documentation
Build integrations
Use it for their own teams
Deploy their own version

The contribution flow is simple:

git clone https://github.com/gavuvapro/prompt-empire.git
cd prompt-empire

Then:

Fork the repository.
Create a new branch.
Add a prompt in the right category.
Follow the prompt format.
Submit a pull request.

The project uses the MIT License, so it is friendly for learning, remixing, and professional use.

Who can use this?

Developers

Use prompts for:

Code reviews
API design
Security audits
Debugging
React refactors
Next.js SEO
Database schema reviews
DevOps incident postmortems
RAG evaluation plans

Designers

Use prompts for:

UI design briefs
UX research interviews
Branding
Logo concepts
Typography systems
Posters and flyers

Founders and businesses

Use prompts for:

Startup idea validation
SaaS pricing
Product requirements documents
Marketing campaigns
Sales discovery calls
Customer support macros

Writers and creators

Use prompts for:

Blog outlines
SEO content briefs
Landing page copy
Storytelling
Newsletters
Technical writing
Video scripts

Students and researchers

Use prompts for:

Active recall
Study planning
Research questions
Concept explanations
Language learning

AI artists and media creators

Use prompts for:

Midjourney
DALL-E
Flux
Stable Diffusion
Leonardo AI
Runway
Veo
Kling
ElevenLabs
Suno
Udio

What makes a good prompt?

One lesson I learned while building this project:

A good prompt is not just a command. It is a structured request.

The best prompts usually include:

A role
A goal
Context
Constraints
Input placeholders
Output format
Evaluation criteria
Examples

Bad prompt:

Write a landing page.

Better prompt:

Act as a senior conversion copywriter.
Write landing page copy for {{product}} targeting {{audience}}.
Include hero headline, subheadline, benefits, objections, social proof, FAQ, and CTA variants.
Use a clear, confident, non-hype tone.

That difference matters.

The first prompt gives you generic output.

The second prompt gives the AI a job, a user, a structure, and a quality bar.