Repolode
jianruntech

jianruntech/geo-score

Free 0–100 score for how well AI answer engines can reach, parse and cite a website, plus optional citation tracking with your own API keys.

Stars
318
Forks
12
Open issues
0
Last commit
yesterday

What is it?

geo-score scores a website from 0 to 100 on whether AI answer engines such as ChatGPT, Perplexity and Gemini can reach, parse, trust and cite it. The score is computed against a published, versioned rubric with 21 tiered checks in five pillars, plus three gates that cap the result when a crawler cannot reach the content. It needs no API key and no installed dependencies beyond Python's standard library.

With your own keys for OpenAI, Perplexity, Gemini, Anthropic or OpenRouter, the tool can also ask engines a question and report which sources the answers cite. A watch mode runs a fixed question set on a schedule and reports citation rates, share of voice and changes between runs. These citation results are shown beside the score and never added to it.

The project is available as a CLI, a GitHub Action, an MCP server for agents and a Claude Code skill. It produces text, JSON, SARIF and JUnit reports.

Key features

  • 01Scores a site 0–100 across reachability, understandability, content citability, brand credibility and answer fit.
  • 02Runs with no API key and only Python's standard library.
  • 03Asks ChatGPT, Perplexity, Gemini and Claude a question and reports the sources cited.
  • 04Watch mode tracks a fixed question set week over week with confidence intervals.
  • 05Ships as a CLI, GitHub Action, MCP server and Claude Code skill.

Use cases

  • Checking whether a company website is reachable and readable by AI search crawlers before a content overhaul.
  • Adding a readiness score check to CI so builds fail when a site regresses.
  • Tracking how often a brand is cited by AI engines against named competitors over time.
  • Letting an AI agent audit a site through the MCP server.

Quick start

  1. 1

    Install the pinned release from PyPI

    pip install geo-score==1.8.3
  2. 2

    Score a website (free, no API key needed)

    geo_score.py stripe.com

Things to know

  • Levels 2 and 3 citation checks need your own search-enabled API keys from each AI engine.
  • Sponsor relays are untested for citation measurements and may not pass web search and citations through unchanged.
  • Scores measure readiness against a rubric the project defines, not a measured effect on AI citations.

Topics

FAQ

What is geo-score?
Free 0–100 score for how well AI answer engines can reach, parse and cite a website, plus optional citation tracking with your own API keys. geo-score scores a website from 0 to 100 on whether AI answer engines such as ChatGPT, Perplexity and Gemini can reach, parse, trust and cite it. The score is computed against a published, versioned rubric with 21 tiered checks in five pillars, plus three gates that cap the result when a crawler cannot reach the content. It needs no API key and no installed dependencies beyond Python's standard library.
Is geo-score free to use? What license does it have?
geo-score is open source under the MIT license. Check the license terms before using it in commercial products.
Is geo-score still maintained?
Actively maintained. The last commit to geo-score was on Oct 19, 2026.
What can I use geo-score for?
Checking whether a company website is reachable and readable by AI search crawlers before a content overhaul. Adding a readiness score check to CI so builds fail when a site regresses. Tracking how often a brand is cited by AI engines against named competitors over time. Letting an AI agent audit a site through the MCP server.
How do I install geo-score?
Install the pinned release from PyPI: pip install geo-score==1.8.3 Score a website (free, no API key needed): geo_score.py stripe.com
Where can I get geo-score?
The source code and installation instructions are on GitHub: https://github.com/jianruntech/geo-score

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects