Skip to content

Repository files navigation

ySEO-PRO-AI

The open-source SEO automation platform that doesn't just audit — it fixes.

11 real MCP tools for technical SEO, auto-fixing, multilingual intelligence, drift monitoring, AI search readiness, and competitive analysis. Works with Claude, Cursor, Codex, Gemini, Windsurf, and any MCP client.

License: MIT MCP Compatible Tests Python


Why ySEO-PRO-AI?

Traditional SEO Tools Audit-Only MCP Servers ySEO-PRO-AI
Audit Reports in dashboard Lists issues Finds issues
Fix Manual work Manual work Auto-fixes
AI Search Not covered Basic AI Readiness Score
Multilingual Basic hreflang check Not covered 14-language intelligence
Monitoring Separate tool Not covered Built-in drift alerts
Programmatic Not covered Not covered City/vs/usecase generator
Indexing Manual submission Not covered IndexNow auto-ping
Cost $100-500/month Free (limited) Free + extensible

Quick Start

Install as MCP Server (Claude / Cursor / Windsurf)

{
  "mcpServers": {
    "yseo-pro-ai": {
      "command": "npx",
      "args": ["-y", "yseo-pro-ai"]
    }
  }
}

Install from Source

git clone https://github.com/cociugv/ySEO-PRO-AI.git
cd ySEO-PRO-AI
npm install

Then ask your AI assistant: "Audit https://example.com for SEO issues and fix them"


MCP Tools (11 real implementations)

Audit

  • seo_audit_page — Full page audit (30+ checks, all modules)
  • seo_check_indexability — Robots, meta robots, canonical, X-Robots-Tag

Auto-Fix

  • seo_fix_auto — Fix all fixable issues (with dry-run support)
  • seo_fix_schema — Auto-detect page type + generate JSON-LD
  • seo_fix_hreflang — Generate complete hreflang tag set

Analysis

  • seo_score_ai_readiness — AI Search Readiness Score (0-100)

Monitoring

  • seo_monitor_baseline — Capture SEO state snapshot
  • seo_monitor_compare — Detect drift vs baseline

Competitor Intelligence

  • seo_competitor_compare — Side-by-side comparison
  • seo_backlink_opportunities — Find link building targets

Publishing

  • seo_indexnow_ping — Notify search engines instantly (batch support)

Architecture

┌─────────────────────────────────────────────────────────┐
│                    MCP Clients                           │
│  Claude Desktop · Cursor · Codex · Gemini · Windsurf    │
└────────────────────────┬────────────────────────────────┘
                         │ stdio (MCP Protocol)
┌────────────────────────▼────────────────────────────────┐
│              TypeScript MCP Server                       │
│         11 tools · Zod validation · Deep adapter         │
└────────────────────────┬────────────────────────────────┘
                         │ subprocess bridge
┌────────────────────────▼────────────────────────────────┐
│              Python Analysis Engine                      │
│                                                         │
│  ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐  │
│  │Inspector │ │ Doctor   │ │Polyglot  │ │Sentinel  │  │
│  │(audit)   │ │(auto-fix)│ │(i18n)    │ │(drift)   │  │
│  └──────────┘ └──────────┘ └──────────┘ └──────────┘  │
│  ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐  │
│  │Architect │ │Publisher │ │ Radar    │ │Citadel   │  │
│  │(pages)   │ │(publish) │ │(compete) │ │(schema)  │  │
│  └──────────┘ └──────────┘ └──────────┘ └──────────┘  │
│                                                         │
│  Pipeline: SCAN → DIAGNOSE → FIX → VERIFY → REPORT     │
└─────────────────────────────────────────────────────────┘

Design Principles

  1. Fix, don't just audit — Every diagnosed issue has an auto-fix path
  2. MCP-native — First-class MCP server, works with any compatible client
  3. Zero-config start — Works out of the box, configure only what you need
  4. Multilingual by default — Built for international sites (14+ languages)
  5. AI-search aware — Optimizes for Google AI Overviews, ChatGPT, Perplexity
  6. Plugin-extensible — Add custom tools, integrations, and data sources
  7. Privacy-first — Runs locally, no data sent to third parties

Platform Compatibility

Platform Integration Setup
Claude Desktop MCP server (stdio) Add to mcp.json
Claude Code MCP + Skills + Agents Plugin auto-discovery
Cursor MCP server Add to .cursor/mcp.json
Kiro CLI Steering rules + MCP Open project in Kiro
Windsurf MCP server Add to configuration
Codex CLI Python CLI python -m src.ops.cli
Gemini Code Assist Python CLI Direct execution
n8n / Zapier HTTP bridge Via MCP HTTP transport
Any MCP client stdio transport Standard MCP protocol

CLI Usage

# Quick page scan
python -m src.ops.cli scan https://example.com

# Full audit with all modules
python -m src.ops.cli audit https://example.com

# Auto-fix (preview)
python -m src.ops.cli fix https://example.com --dry-run

# Auto-fix (apply)
python -m src.ops.cli fix https://example.com

# AI Search Readiness Score
python -m src.ops.cli ai-score https://example.com

Configuration

Edit config/default.yaml:

targets:
  primary_domain: "example.com"
  languages: [en, de, fr, es, ro]

modules:
  doctor:
    auto_fix: true
    dry_run: false
  sentinel:
    snapshot_interval_hours: 24
  publisher:
    indexnow_key: "your-key"

Environment variables (.env):

INDEXNOW_KEY=your-key
PAGESPEED_API_KEY=optional
GSC_CREDENTIALS_JSON=optional

Modules

Module Purpose Key Feature
Inspector Technical SEO scanner 30+ diagnostic checks
Doctor Auto-fix engine Generates fixes, not just reports
Polyglot Multilingual SEO Hreflang audit + generation
Sentinel Drift monitoring Snapshot comparison + alerts
Architect Programmatic SEO City/comparison/usecase pages
Publisher Content publishing Blog API + IndexNow ping
Radar Competitor intel Compare + backlink opportunities
Citadel Schema + AI readiness Auto-inject + scoring

Testing

# Run full test suite (113 tests)
python -m pytest tests/ -v

# Run only benchmark fixtures
python -m pytest tests/test_benchmark.py -v

# Run specific module tests
python -m pytest tests/test_remediation.py -v

Test coverage:

  • Parser: 21 tests (extraction, broken HTML, serialization)
  • Pipeline: 13 tests (stages, execution, scoring, hooks)
  • Doctor: 9 tests (fix status honesty, generation, artifacts)
  • Modules: 22 tests (hreflang, drift, architect, radar, citadel)
  • Operations: 8 tests (deep module, typed records)
  • Remediation: 18 tests (state machine, file adapter, integration)
  • Benchmark: 22 tests (precision/recall against HTML fixtures)

Contributing

Contributions welcome! See CONTRIBUTING.md for guidelines.

Key areas:

  • New MCP tools
  • Additional fix prescriptions
  • Integration plugins (Ahrefs, Semrush, etc.)
  • Translations for multilingual content templates
  • Documentation improvements

License

MIT License — Vadim Cociug


Links

About

No description or website provided.

Topics

Resources

Contributing

Security policy

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages