Three months ago, I launched CitableHub a platform designed to make software projects discoverable and citable by AI search engines like ChatGPT, Perplexity, Gemini, and Claude.
The idea was simple but radical: traditional SEO optimizes for Google's index. But AI models don't use Google's index. They scrape, parse, and synthesize differently. So I built a platform specifically engineered for what I call Generative Engine Optimization (GEO) making your project easy for AI to find, understand, verify, and cite.
Yesterday, Google's Gemini Deep Research their autonomous research agent — decided to analyze CitableHub on its own.
It scraped the site. Read the profiles. Compared it to Product Hunt, Futurepedia, and SaaSHub. Then wrote a 3,000+ word analytical report.
And it understood everything.
What Gemini Found
Gemini described CitableHub as:
"An entity resolution platform engineered for AI search systems."
That's literally the positioning I built it around — but I never wrote that sentence anywhere on the site. Gemini inferred it from the structured data.
Here's what it correctly identified:
The 5-stage pipeline for project optimization
The Citability Score with its 5 dimensions: Identity (25%), Evidence (30%), Trust (20%), Freshness (10%), Classification (15%)
GQI (Generative Query Impressions) and how GQI Boost works
Academic citations — BibTeX, APA, MLA formats for every project
Verification IDs (CH-VER-XXXXXX) as trust signals
The white-label partner engine at 0% commission
It even analyzed real profiles with their scores:
Song Finder: 51/100 (CH-VER-437136)
Makify AI: 63/100 "Strong" (CH-VER-472772)
Descript: 52/100 (CH-VER-556825)
Standard Notes: ~75/100
The Comparison Table That Made My Day
Gemini built a comparison matrix between CitableHub and traditional directories:
Feature CitableHub Product Hunt Futurepedia
Target audience AI engines Humans Humans
Data format JSON-LD + Schema.org HTML HTML
Ranking metric Citability Score Upvotes Reviews
Listing permanence Permanent 24-48h window Permanent
Structured for LLMs ✅ Native ❌ Not designed ❌ Not designed
This is the exact differentiation I've been articulating for months. Seeing an AI model independently arrive at the same conclusion — and present it more clearly than most humans could — was surreal.
Why This Matters for Developers
If you're building a dev tool, an API, or any software product, here's the uncomfortable truth:
When someone asks ChatGPT "what's the best tool for X?", your project needs to be in that answer.
Traditional directories optimize for human eyeballs. But AI models don't browse Product Hunt. They parse structured data, verify claims against evidence, and synthesize answers from sources they can trust.
That's what GEO solves:
Structured identity Schema.org markup, JSON-LD, canonical metadata that AI can parse without guessing
Evidence layer Proof of claims (benchmarks, testimonials, integrations) that models can verify
Trust signals Verification IDs, academic citations, freshness timestamps
Classification Proper taxonomic categorization so AI knows what category your tool belongs to
Freshness Active signals that tell crawlers "this project is alive and maintained"
Gemini cited 14 pages from citablehub.com in its report. That's 14 URLs that an AI research agent deemed authoritative enough to reference. For a 3-month-old platform.
The Technical Bit
For those curious about how it works under the hood:
Every project profile generates structured data (Schema.org SoftwareApplication + FAQPage) that AI crawlers can parse natively
An llms.txt endpoint serves machine-optimized project data
The Citability Score is computed across 5 weighted dimensions — not vanity metrics, but signals that map to how LLMs evaluate source reliability
We track which AI crawlers visit (DeepSeek, Perplexity, Claude, GPTBot, Gemini, and others) real data, not estimates
The fact that Gemini's autonomous research agent could read, understand, and accurately report on all of this without any human explanation is proof that the structured data approach works.
The Takeaway
We're in a transition period. SEO isn't dead, but the discovery layer is changing. AI models are becoming the first point of contact between users and products.
If your project isn't structured for AI retrieval, you're invisible to a growing percentage of your potential users.
CitableHub is free to list. 700+ projects are already there. And apparently, Google's own AI thinks it's worth writing a research paper about.
Have you checked if AI models can find and accurately describe your project? Try asking ChatGPT or Perplexity about it — you might be surprised by what they say (or don't say).
Top comments (0)