Open-source tools for the AI era.
We don't build wrappers. We build infrastructure.
Search is being eaten by AI. Websites optimized for Google still don't exist for ChatGPT, Perplexity, or Claude. We're fixing that.
One vertical. One mission: make websites visible to AI search engines.
GEO vertical → Audit, optimize, and track AI citation readiness
SaaS layer → Monitoring, history, and team features on top of the engine
"If an AI can't cite your website, you don't exist."
geo-optimizer-skill — The engine
The first open-source GEO audit and optimization engine. Based on the Princeton KDD 2024 research paper + AutoGEO ICLR 2026 methods.
15 CLI commands · 8 scoring categories · 47 research-backed methods · 7 output formats · 1,720 tests · MIT licensed · PyPI published · MCP server compatible · Astro integration ready.
Technical deep dive
Scoring engine — Weighted 0–100 score across 8 categories:
| Category | What it checks |
|---|---|
robots.txt |
AI bot access (GPTBot, ClaudeBot, PerplexityBot, etc.) |
llms.txt |
AI-readable site index — the new sitemap.xml for LLMs |
JSON-LD Schema |
Structured data richness and validity |
Meta tags |
SEO signals that AI engines actually parse |
Content quality |
Heading hierarchy, content depth, readability |
AI signals |
Citations, authoritative language, E-E-A-T markers |
AI discovery |
Discoverability across ChatGPT, Perplexity, Claude, Gemini |
Brand coherence |
Consistent entity representation across the site |
MCP integration — 11 tools exposed via FastMCP:
geo_audit · geo_fix · geo_llms_generate · geo_citability · geo_schema_validate · geo_compare · geo_gap_analysis · geo_ai_discovery · geo_check_bots · geo_trust_score · geo_negative_signals · geo_factual_accuracy
Security — Anti-SSRF hardened. All URLs validated via DNS pinning. Streaming responses with 10MB size limit. No direct requests.get() — ever.
47 citability methods derived from Princeton + ICLR research for improving AI citation probability.
Bonus checks — CDN crawler access, JS rendering, WebMCP readiness, negative signals (8 anti-citation patterns), prompt injection detection, trust stack score, RAG chunk readiness, content decay prediction, platform citation profiles, multimodal readiness.
geoready-platform — The SaaS
Live at geoready.dev — authenticated SaaS with dashboard, audit history, domain monitoring, citation tracking, drift alerts, and billing. Built on top of the open-source engine.
wordpress-geoready — The plugin
WordPress plugin for AI visibility & Answer Engine Optimization. Generates llms.txt + AI discovery files and scores AI search readiness. Powered by the open-source GEO Optimizer engine.
graph LR
A["geo-optimizer-skill<br/>(open-source engine)"] -->|powers| B["geoready-platform<br/>(SaaS)"]
A -->|powers| C["wordpress-geoready<br/>(plugin)"]
B -->|monitors| D["geoready.dev<br/>(live site)"]
style A fill:#0c2d48,stroke:#0ea5e9,color:#fff
style B fill:#0c2d48,stroke:#0ea5e9,color:#fff
style C fill:#0c2d48,stroke:#0ea5e9,color:#fff
style D fill:#0c2d48,stroke:#0ea5e9,color:#fff
- Local first — Your data stays on your machine. Zero telemetry, zero cloud dependencies in the CLI.
- Research-backed — We implement peer-reviewed methods (Princeton KDD, ICLR), not blog post hype.
- MCP native — The engine speaks Model Context Protocol. Plug into any AI assistant.
- Production grade — 1,720 tests, SSRF protection, CI/CD integration, SARIF + JUnit output. Not demos — infrastructure.