A curated library of token-efficient prompts — for developers, creators, educators, and everyone.
Every prompt ships with Lean, Balanced, and Max Quality variants.
PromptLean is a curated, open-source library of 120 prompts across 14 categories — from code review and SQL optimization to creative writing, journaling, interview prep, and emotional support.
The key idea: most prompts are overengineered. You don't need 200 tokens of preamble for a code review. But sometimes you do need the full rubric. PromptLean gives you the right prompt for the right moment.
Every prompt ships with three variants:
| Variant | Typical tokens | When to use |
|---|---|---|
| ⚡ Lean | 18–40 | Fast iteration, tight context windows, high-volume API calls |
| ⚖ Balanced | 50–110 | Structured output with consistent format across runs |
| ★ Max Quality | 85–230 | External deliverables, research-grade tasks, expert review |
The Lean variant saves 80–87% of tokens vs Max Quality with surprisingly little quality loss for well-understood tasks.
Every prompt detail page also includes a Model × Variant comparison table — quality scores (1–5) across 8 frontier models with recommended variant per model.
- ⌨ Playground — fill
[PLACEHOLDER]slots with live preview and one-click copy of the completed prompt - ⊞ Diff view — word-level diff between Max Quality and Lean, side by side with additions/removals highlighted
- Cost estimator — sidebar dropdown: pick a model, see estimated cost per 1,000 API calls for each variant
- ★ Favorites — star any prompt to save it; browse page has a "Saved" filter
- 🔥 Popular badge — appears on cards after 10+ copy-clicks (tracked in
localStorage)
- Token budget filters — ⚡ < 30t · ⚖ 30–100t · ★ 100+t filter chips on the lean variant
- Saved filter — show only starred prompts
- Tag + category filters — sidebar with all 14 categories and top 18 tags
- Workflow stepper — 7 predefined workflows (Code Quality, Writing, Research, Career, Data, Learning, Engineering); shows where the current prompt sits and links to the next step
- Community variant tabs — extra tabs loaded from
community/{id}.json; fuchsia-styled to distinguish from standard variants
All prompts are available as a raw JSON endpoint — no key, no auth:
https://raw.githubusercontent.com/kishormorol/promptlean/main/data/prompts.json
const { prompts } = await fetch(
'https://raw.githubusercontent.com/kishormorol/promptlean/main/data/prompts.json'
).then(r => r.json());CORS is open (access-control-allow-origin: *). Full docs on the About page.
20 new prompts added in the latest expansion (100 → 120):
| Category | New Prompts |
|---|---|
| AI / ML | Fine-Tuning Dataset Designer · Prompt Chain Architect · RAG Query Optimizer · AI Feature Product Spec · Model Selection Advisor |
| Entertainment & Creative | Game Narrative Designer · Worldbuilding Assistant · Screenplay Scene Writer · Trivia Question Generator · Comedy Sketch Writer |
| Safety & Crisis Support | Boundary-Setting Script · Personal Safety Plan Builder · Conflict De-escalation Guide |
| Research | Literature Review Outliner |
| Data | SQL Query Optimizer |
| Career & Professional Growth | Tailored Cover Letter Writer |
| Education & Tutoring | Flashcard Generator |
| Engineering | API Documentation Writer |
| Well-being & Mental Health | Daily Reflection Coach |
| General Purpose | Problem Reframing Toolkit |
120 prompts across 14 categories:
| Category | Count | Prompts |
|---|---|---|
| Code | 10 | Code Review · Explain Code to a Junior Dev · Write Unit Tests · Debug & Fix a Bug · Refactor Code · Write Code Documentation · Generate a Regular Expression · Translate Code to Another Language · Developer Relations Consultant · Cybersecurity Strategy Advisor |
| Data | 9 | Optimize a SQL Query · Data Cleaning Plan · Exploratory Data Analysis Plan · Statistical Analysis Advisor · Annual Report Analyst (10-K) · Excel Formula Specialist · Unstructured Text to JSON · Data Visualization Advisor · SQL Query Optimizer |
| Writing | 6 | Write a Technical Blog Post · Write a Cold Email · Write an Executive Summary · Ad Copy Generator · Social Media Post Optimizer · FAQ Generator |
| Productivity | 7 | Meeting Notes → Action Items · Weekly Planner & Prioritization · Draft a Professional Email Reply · Decision Framework · Startup Idea Generator · Personal Finance Advisor · Legal Contract Reviewer |
| Engineering | 8 | System Design Review · Write a Product Requirements Document · Incident Postmortem · Architecture Decision Record (ADR) · IT Solution Architect · UX/UI Design Reviewer · Software Project Planner · API Documentation Writer |
| Research | 5 | Summarize a Research Paper · Literature Review Synthesis · Debate Both Sides · Venture Capital Analyst · Literature Review Outliner |
| Education & Tutoring | 7 | Explain Any Concept · Socratic Tutoring Session · Generate Practice Questions · Lesson Plan Generator · Study Guide Creator · Word Etymology Explorer · Flashcard Generator |
| Career & Professional Growth | 7 | Resume Review · Interview Preparation Coach · Write a LinkedIn Post · Salary Negotiation Script · Performance Review Self-Assessment · Negotiation Coach · Tailored Cover Letter Writer |
| General Purpose | 7 | Summarize Any Text · Brainstorm Ideas · Explain Like I'm 5 (ELI5) · Give Constructive Feedback · Fallacy Finder · Product Critique & Feedback · Problem Reframing Toolkit |
| AI / ML | 7 | LLM Evaluation Rubric · Write a System Prompt · Prompt Optimizer · System Prompt Debugger · Fine-Tuning Dataset Designer · Prompt Chain Architect · RAG Query Optimizer · AI Feature Product Spec · Model Selection Advisor |
| Well-being & Mental Health | 4 | Daily Reflection & Journaling Guide · Stress Relief & Reframing · Goal Setting & Action Plan · Daily Reflection Coach |
| Entertainment & Creative | 8 | Short Story Generator · Poem Generator · Character Creator · Game Narrative Designer · Worldbuilding Assistant · Screenplay Scene Writer · Trivia Question Generator · Comedy Sketch Writer |
| Safety & Crisis Support | 4 | Compassionate Crisis Response · Boundary-Setting Script · Personal Safety Plan Builder · Conflict De-escalation Guide |
| Supportive Conversation | 1 | Active Listening & Emotional Support |
All prompts include notes and quality benchmarks for 8 frontier models:
| Model | Provider |
|---|---|
gpt-5.4 |
OpenAI |
claude-sonnet-4-6 |
Anthropic |
claude-opus-4-6 |
Anthropic |
gemini-3.1-pro |
|
grok-4 |
xAI |
llama-4-scout |
Meta (open weight) |
mistral-large-3 |
Mistral (open weight) |
o1 |
OpenAI |
Each prompt detail page shows a Model × Variant matrix with color-coded quality scores and a recommended variant badge per model.
| Page | URL |
|---|---|
| Home | / |
| Browse | /browse.html — search, filter by category/tag/budget/saved |
| Prompt detail | /prompt.html?id=... — variants, playground, diff, cost estimator, workflow, community tabs |
| Compare | /compare.html — 3-column side-by-side view |
| About | /about.html — includes API docs |
| Contribute | /contribute.html — prompt builder + community variant guide |
| Tool | Free | Open source | Token variants | Model benchmarks | 14+ categories |
|---|---|---|---|---|---|
| PromptLean | ✅ | ✅ | ✅ | ✅ | ✅ |
| AwesomePrompts | ✅ | ✅ | ❌ | ❌ | ❌ |
| FlowGPT | Partial | ❌ | ❌ | ❌ | ✅ |
| PromptHero | Partial | ❌ | ❌ | ❌ | ✅ |
| LangChain Hub | ✅ | ✅ | ❌ | ❌ | ✅ |
PromptLean is the only prompt library that gives you three token-optimized variants per prompt with quality benchmarks across 8 frontier models — free, no sign-up, open source.
All prompts live in data/prompts.json. Use the Prompt Builder on the contribute page to generate the correct JSON, then open a PR.
Schema:
{
"id": "your-prompt-id",
"title": "Your Prompt Title",
"category": "Code",
"tags": ["tag1", "tag2"],
"featured": false,
"description": "One sentence describing what this prompt does and when to use it.",
"model_notes": {
"gpt-5.4": "...",
"claude-sonnet-4-6": "...",
"claude-opus-4-6": "...",
"gemini-3.1-pro": "...",
"grok-4": "...",
"llama-4-scout": "...",
"mistral-large-3": "...",
"o1": "..."
},
"benchmarks": {
"gpt-5.4": { "lean": 4, "balanced": 5, "max_quality": 5, "best": "balanced" },
"claude-sonnet-4-6": { "lean": 5, "balanced": 5, "max_quality": 4, "best": "lean" }
},
"variants": {
"lean": { "prompt": "...", "token_estimate": 24 },
"balanced": { "prompt": "...", "token_estimate": 72 },
"max_quality": { "prompt": "...", "token_estimate": 160 }
}
}Have a specialized take on an existing prompt? Create community/{prompt-id}.json:
[
{
"id": "your-variant-id",
"label": "🔒 Your Variant Name",
"author": "your-github-username",
"description": "One sentence: what does this variant optimize for?",
"prompt": "Your prompt text. Use [PLACEHOLDER] for user content.",
"token_estimate": 95
}
]Community variants appear as extra tabs on the prompt detail page. See community/code-review.json for a working example.
Full contribution guidelines: contribute.html
- Pure HTML + CSS + Vanilla JS — no framework, no build step
- Data-driven — all prompts in
data/prompts.json, community variants incommunity/ - GitHub Pages — deployed via GitHub Actions on every push to
main - Dark mode default, light mode toggle, persisted in
localStorage
git clone https://github.com/kishormorol/promptlean.git
cd promptlean
# open index.html in a browser, or serve with any static server:
python -m http.server 8000
# then visit http://localhost:8000Some prompts are adapted from open-source collections. All adaptations are significantly rewritten into PromptLean's three-tier format with model notes and benchmarks added. Source metadata is stored in data/prompts.json for full traceability.
| Source | License | Prompts adapted |
|---|---|---|
| awesome-chatgpt-prompts by Fatih Kadir Akın | CC0 1.0 (public domain) | 11 |
| LLM-Prompt-Library by Alexander Bilzerian | MIT © 2025 Alexander Bilzerian | 7 |
| Community Gist by BigDog1400 | MIT | 1 |
Full attribution details: ATTRIBUTION.md
MIT — use freely, contribute back.