Here’s the gap in the market:
Your clients have an SEO problem. They don’t know which keywords are worth ranking for. Their competitors are stealing traffic from 200+ keywords they’ve never targeted. And hiring an SEO specialist costs $3,000-$5,000 a month or $150/hour for freelancers.
So they stay stuck. Guessing on keywords. Writing headlines that don’t convert. Watching competitors dominate the SERPs.
You already know this problem exists. What you need is a way to solve it for them, fast, repeatably, and profitably.
The Unfair SEO Agent does this. In 2 minutes, it finds 200 keyword opportunities. In 30 minutes, it generates and evolves 50+ headlines to 85+ quality. Everything is done. No research. No guessing. No hiring.
And because it comes in two versions (CLI for technical users and Desktop for non-technical users), you can sell it, white-label it, integrate it into your service offerings, or deliver it as part of your AI automation package.
WHAT IS IT?
The Unfair SEO Agent is a two-edition AI system that automates three core SEO tasks:
- FIND GAPS Input: 2-3 competitor domains Output: 200 ranked keyword opportunities with search volume and difficulty Speed: 30 seconds Uses: Claude Sonnet 4 + intelligent gap analysis prompt
- GENERATE HEADLINES Input: A target keyword Output: 50 unique SEO headlines using different approaches (data-driven, question-based, listicle, curiosity, benefit, comparison) Speed: 15 seconds Uses: Claude Sonnet 4 + multi-angle generation prompt
- SCORE & EVOLVE Input: Your 50 headlines Output: Automated evolution through 3 rounds. Headlines scored by a panel of 4 experts (CRO specialist, SEO expert, copywriter, skeptical reader) across 4 dimensions (clickability, SEO strength, clarity, trust). Final winner: 85+ quality score. Speed: 60 seconds per round (3 rounds = 2-3 minutes) Uses: Claude Sonnet 4 + multi-expert scoring system
The entire workflow takes 2-5 minutes to run. Each keyword is production-ready and tested against multiple quality criteria.
HOW IT WORKS
EDITION 1: CLI (For You, For Your Developer Clients, For Automation)
Three bash commands. That’s it.
python unfair-seo-cli.py find-gaps competitor1.com competitor2.com
→ Returns: gaps.json (200 keywords ranked by opportunity)
python unfair-seo-cli.py generate-variants "your keyword" --count 50
→ Returns: variants.json (50 unique headlines)
python unfair-seo-cli.py evolve variants.json --rounds 3 --target-score 85
→ Returns: evolution-result.json (evolution history + winning headline)
Single Python file. No dependencies except anthropic SDK + python-dotenv. Works in Claude Code, your terminal, or integrated into any automation pipeline (Make, n8n, Zapier, Activepieces).
Setup: bash setup.sh → add ANTHROPIC_API_KEY → done. Under 2 minutes.
EDITION 2: DESKTOP (For Your Non-Technical Clients, For Manual Use)
Four copy-paste prompts. No code. Works in the Claude.ai or the Claude Desktop app.
WORKFLOW:
- Open Prompt 1 (find_competitor_gaps.txt)
- Paste into Claude with competitor URLs
- Get back 200 keywords, pick your top 5
- For each keyword: Prompt 2 → generates 50 headlines
- Paste headlines into Prompt 3 → scores all 50
- If top score < 80, paste Prompt 4 → evolves them further
- Deploy your winner
Speed: 5 minutes of research, 30 minutes per keyword (with evolution). A non-technical user can do this solo. No training needed.
Setup: Unzip, read START_HERE.txt, copy the first prompt, paste into Claude. Done.
WHO IS THIS FOR?
PERSONA 1: AI AUTOMATORS
You build AI agents and automation workflows for clients. You integrate tools like Claude, Make, n8n, Activepieces, and Zapier. You charge by the workflow, by the month, or by results.
This agent is your next profitable offer.
Why: You can deliver SEO keyword research and headline generation as a standalone service. $297-$497 per use, or $297/month for recurring access. Deploy it via API, integrate it into your client’s dashboard, white-label it, or include it in a larger automation package.
Time to revenue: 2 hours (rebrand, add to offer page, start selling)
Example use case: “Your agency needs keyword research + landing page headlines for 5 new client verticals. Instead of 40 hours of manual work, you deploy the CLI version, run it 5 times, get 1,000 ranked keywords and 250 winning headlines back. Done in 30 minutes.”
PERSONA 2: AI AGENT BUILDERS
You build and sell custom AI agents. You package them as products, sell them on Gumroad, use them as lead magnets, and integrate them into larger SaaS offerings.
This agent is a pre-built, tested, revenue-generating product you can customize immediately.
Why: The agent is production-ready. All you do is rebrand it, change the prompts to match your style, and sell it. No development. No testing. Two editions mean two different price points (CLI = $197, Desktop = $97, or bundle at $247).
Time to revenue: 1 hour (rebrand, upload, launch)
Example use case: “You’re running an AI skills marketplace. You add the Unfair SEO Agent as a premium offering. Customers buy it, download the ZIPs, and use them immediately. You get 100% profit margin on the first month, then handle support.”
PERSONA 3: AGENCIES
You run a content, marketing, or SEO agency. You need a faster way to conduct keyword research and generate headlines for clients.
This agent is your competitive moat.
Why: You deploy the Desktop Edition for non-technical clients (they can use it themselves or you run it for them). You deploy the CLI Edition for technical clients or internal use. Every keyword research project that used to take 20 hours now takes 30 minutes. You either deliver faster, charge the same, and keep the difference, or deliver the same speed but take on 10x more clients.
Time to revenue: Immediate (start using it on the next client project)
Example use case:
“Your client is an e-commerce store. You run the agent on their 3 top competitors. 200 keywords back. You pick the top 20 that fit their niche. For 10 of those, you generate headlines + content. Client launches 10 new blog posts in 2 weeks instead of 2 months. You charge $2,000-$5,000 for the research + content strategy. Cost to you: 4 hours of your time.”
WHY IT WORKS
REASON 1: SEO IS A NUMBERS GAME. THIS GIVES YOU THE RIGHT NUMBERS.
99% of SEO strategies fail because they target the wrong keywords.
Competitors rank for 200+ keywords you haven’t even heard of. Those keywords are converting. They’re getting traffic. And you’re leaving money on the table by not targeting them.
The Unfair SEO Agent finds these keywords automatically. Not guesses. Not “keywords you think are important.” Real keywords your competitors rank for, with search volume, difficulty, and opportunity scores baked in.
Result: Your clients target the right 20 keywords instead of wasting 6 months on keywords that have no search volume.
REASON 2: HEADLINES DETERMINE CLICK-THROUGH RATE. EVOLUTION DETERMINES QUALITY.
A bad headline means no clicks, which means no traffic, no leads, and no sales.
Most people write ONE headline and hope it works. The agent generates 50 headlines using proven copywriting approaches: data-driven hooks, curiosity gaps, listicles, benefit statements, and comparisons.
Then it scores them through a multi-expert panel (CRO specialist, SEO expert, copywriter, skeptical reader). Evolution happens automatically. Round 1 scores the best. Round 2 improves on the winners. Round 3 polishes to perfection.
Result: Your clients go from 1 headline (maybe 65/100 quality) to 1 winner (85+/100 quality). That 20-point difference is 30-40% more clicks.
REASON 3: SPEED = PROFIT.
Keyword research + headline generation used to take 20-40 hours per project.
The agent does it in 2-5 minutes.
If you’re selling this as a service, that’s $3,000-$5,000 of revenue per 2 minutes of work. Margin is near-infinite.
If you’re white-labeling it: Your customers get results faster. They’re happier. They refer more people. Your Net Promoter Score goes up.
If you’re including it in a bigger package, you can take on more clients because your turnaround time has dropped by 90%.
REASON 4: TWO EDITIONS = TWO MARKETS = MORE REVENUE.
Most AI tools are either for developers (CLI) or non-technical users (web UI). You have to pick.
The Unfair SEO Agent comes in both. One codebase (technically), two different interfaces.
CLI Edition: Sells to developers, agencies, automators, and folks who want to integrate it into workflows.
Desktop Edition: Sells to non-technical founders, solopreneurs, content teams, folks who want to paste prompts into Claude.
Same product. Different distribution channels. Different price points. Different customer personas.
If you’re reselling: Offer both. Some customers will buy CLI, some will buy Desktop, and some will buy the bundle ($247). You hit all three.
REASON 5: ZERO CONFIGURATION FRICTION.
Other SEO tools make you:
- Connect API keys to 5 different services
- Set up complex configurations
- Learn platform-specific UI
- Wait 2 weeks for training
This agent:
- CLI: bash setup.sh, add ANTHROPIC_API_KEY, done
- Desktop: Unzip, read START_HERE.txt, paste prompts, done
Setup time: 2 minutes max. Customer confusion: near zero.
Result: Faster customer activation. Faster time-to-value. Better reviews. More referrals.
WHAT’S INCLUDED
CLI EDITION
- unfair-seo-cli.py (single Python file, 600 lines, production-ready)
- setup.sh (automated installation)
- requirements.txt (anthropic + python-dotenv only)
- .env.example (environment template)
- AGENTS.md (behavior specification)
- README.md (full documentation with examples)
- └── [Ready to customize, rebrand, integrate, or resell]
DESKTOP EDITION
- START_HERE.txt (quick start guide)
- unfair-seo-skill.md (complete skill file with all prompts + interpretations)
- Connector_Guide.md (technical reference)
- Prompts/ (4 copy-paste prompts) │ 01_find_competitor_gaps.txt │ 02_generate_headline_variants.txt │ 03_score_and_evolve.txt │ 04_publish_winner.txt
- Knowledge_Base/ (reference materials) │competitor_urls_template.txt │seo_scoring_rubric.txt
- [Ready to customize, rebrand, share with clients]
BOTH EDITIONS
✓ Production-ready. No missing files, no broken code.
✓ Tested. Python syntax validated, JSON I/O tested, workflows verified.
✓ Documented. Inline comments, README, guides, and FAQs included.
✓ Customizable. Rebrand, modify prompts, change scoring criteria, and adjust API calls.
✓ Integrated. Works with the Claude API (no other dependencies for the Desktop edition).