GEO‑SEO Claude tool released
A free open-source tool called GEO‑SEO Claude has been published to audit websites for AI-driven search visibility and generate actionable reports to boost citations in services like ChatGPT and Perplexity. Built on Claude, it produces audits and proposed fixes to improve how sites surface in AI search results, which is relevant for web apps relying on discoverability. That offers a hands-on way to test and adapt content for the emerging AI search layer. (x.com)
A new open-source project called geo-seo-claude is trying to do for artificial intelligence search what old-school search engine optimization did for Google: inspect a site, score what machines can actually use, and hand back a repair list. The GitHub repository says it audits “citability,” artificial intelligence crawler access, schema markup, brand authority, and exports reports, and it was sitting at about 4,900 GitHub stars when I checked on April 11, 2026. (github.com) The basic bet is that more people now ask ChatGPT, Perplexity, Gemini, and Google AI Overviews for answers instead of clicking ten blue links. The project’s README explicitly targets those products and says it is built for “AI-powered search engines” rather than only traditional search rankings. (github.com) That changes what “being found” looks like. A webpage used to win by ranking high for a keyword, but an answer engine often wins by pulling one clean paragraph, one table, or one fact block it can quote back to the user. (github.com) Geo-seo-claude is packaged as a skill for Claude Code, which is Anthropic’s command-line coding environment. The repo’s install instructions use shell scripts for macOS, Linux, and Windows, then run commands like `/geo audit` against a site URL. (github.com) The tool’s pitch is not “write better marketing copy.” It is closer to a building inspector checking whether robots can enter the house, whether the rooms are labeled, and whether the important facts are easy to pick up without reading the whole thing. (github.com) One of its core ideas is “citability scoring,” which means grading whether a page contains self-contained passages an answer engine can safely quote. The repository also lists checks for crawler access and `llms.txt`, which is a proposed file publishers use to tell large language model systems where useful content lives. (github.com) It also leans on schema markup, which is the hidden label system websites use to tell machines “this is a product,” “this is a review,” or “this is an organization.” If a page is a grocery shelf, schema markup is the barcode that helps software know what each item is without guessing from the packaging. (github.com) The project is arriving into a small boom of terminal-based audit tools built around Claude Code. A separate open-source project called Claude SEO advertises 14 sub-skills, 9 parallel agents, and a dedicated `/seo geo` command for Google AI Overviews, ChatGPT, Perplexity, and Bing Copilot, which shows this is already turning into a software category rather than a one-off script. (claude-seo.md) That is why this release is getting attention from web app teams, agencies, and content publishers. If traffic starts upstream in an answer box instead of on a search results page, then the site that gets cited becomes the site that gets the click, and tools like this are trying to make that process less mysterious. (github.com) The catch is that no audit tool can force ChatGPT or Perplexity to cite a page. What it can do is raise the odds by fixing machine-readable structure, exposing crawlable content, and turning vague pages into pages with extractable facts, which is exactly the checklist geo-seo-claude is built to generate. (github.com)
