Skip to content
E-GEO

Getting Started

E-GEO runs in two ways: as a standalone Python CLI (egeo) or as a set of Claude Code agents and skills. Both share the same optimization logic — there is one source of truth.

Prerequisites

Section titled “Prerequisites”
  • Python 3 for the standalone CLI, or Claude Code for the agent workflow
  • A content file or website URL you want to optimize
  • Optional: an OpenAI-compatible API key for real model runs (GEO_EVAL_MOCK=1 works without one)

Installation

Section titled “Installation”

Option 1: Standalone Python CLI

Section titled “Option 1: Standalone Python CLI”
Terminal window
git clone https://github.com/mverab/eGEOagents.git
cd eGEOagents
pip install -e .

Or run without installing:

Terminal window
pip install pyyaml jsonschema
python -m egeo --help

Option 2: Claude Code skills (skills.sh)

Section titled “Option 2: Claude Code skills (skills.sh)”
Terminal window
npx skills add https://github.com/mverab/eGEOagents

Install a single skill:

Terminal window
npx skills add https://github.com/mverab/eGEOagents --skill competitive-analysis

Available skills: competitive-analysis, content-scoring, schema-generator, validation-doctor, geo-loop. Collection page: skills.sh/mverab/egeoagents.

Option 3: Copy the .claude/ folder

Section titled “Option 3: Copy the .claude/ folder”
Terminal window
git clone https://github.com/mverab/eGEOagents.git
cp -r eGEOagents/.claude /path/to/your/project

First optimization

Section titled “First optimization”

With the CLI

Section titled “With the CLI”
Terminal window
# Deterministic offline run — no API key needed
GEO_EVAL_MOCK=1 egeo optimize examples/sample-input.md --out-dir ./geo-output


# Real run (requires OPENAI_API_KEY)
egeo optimize path/to/page.md --query "best geo tool" --schema-type Article

With Claude Code

Section titled “With Claude Code”
/output-style geo-optimizer
/geo https://yoursite.com

What you get

Section titled “What you get”

The CLI path writes an optimization package to geo-output/ (the Claude Code agent path additionally produces a step-by-step checklist.md):

geo-output/
├── report.md           # Executive summary with scores
├── analysis.json       # Raw analysis data
├── optimized/
│   └── yoursite.md     # Rewritten, GEO-optimized content
├── schema/
│   └── yoursite.json   # JSON-LD schema markup
└── checklist.md        # Step-by-step implementation guide

Copy the optimized content and schema to your website.

What happens during optimization

Section titled “What happens during optimization”

E-GEO runs 4 specialized agents in sequence:

Agent What it does Output
Analyzer Extracts content, scores GEO signals, finds gaps analysis.json
Ranker Simulates AI-engine ranking, predicts position Baseline score
Rewriter Optimizes content while preserving brand voice optimized/*.md
Indexer Generates schema markup and technical assets schema/*.json

Next steps

Section titled “Next steps”