RankMeFast is a self-hosted, AGPL-3.0 SEO platform that reached 105 GitHub stars in a week

Every SEO team runs into the same choice: the software that audits the site also stores every audit, every keyword list, and every client's crawl on someone else's server, and charges per seat. RankMeFast is a new open-source project that rearranges the deal. It runs on your own machine, pays the data vendors directly, and keeps every finding in your own database.

The repository at SelmiAbderrahim/rankme.fast was created on 22 September 2026 and reached 105 stars in the seven days that followed, per the GitHub API on 28 September. It is written in TypeScript, licensed AGPL-3.0, and packaged as a single Docker Compose stack for a team to run itself.

What it does

The README lists site audits (26 automated checks, sorted into fix now, watch, and passed), rank tracking on Google, Bing, YouTube and Amazon, keyword research with SERP-overlap clustering and cannibalization reports, backlink profiles with a link-gap tool, Search Console integration for performance and URL Inspection, and an AI visibility section that tracks prompts in AI Overviews and LLM answers with a weekly digest. Reports come out as PDF or CSV, and a read-only MCP server is included for AI assistants to query the results.

Local and app SEO are covered, seven languages ship with the UI (English, Arabic with right-to-left layout, French, German, Spanish, Russian, Chinese), and each vendor-backed feature has its own on/off setting so a team turns on only the parts it has API keys for.

How the data trade works

The crawl and the SERP data come from third-party APIs called with the team's own keys, and RankMeFast passes the vendor cost through with no markup. The list in the README is DataForSEO, the Google APIs (Search Console, GA4, PageSpeed Insights, CrUX), Firecrawl, and optionally an LLM provider. Data lands in MongoDB and PostgreSQL that the team runs. Every vendor also has a fake backend that returns fixed demo data, so a team can try the whole product locally before paying any vendor.

The trade is real, and worth naming. A closed SaaS suite pays for its own crawl and puts that cost inside the seat price, so the pricing is simple and the data belongs to the vendor. RankMeFast asks a team to pay the crawl and data vendors itself, in exchange for owning the crawl, the keys, the storage, and the exports.

Where the pitch stops

The project is at version 0.1.0 and marked public beta by the maintainer, with breaking changes expected between releases and bug reports collected on GitHub issues. Traction is real (105 stars in seven days, 7 forks, verified 28 September 2026 through the GitHub API) and the project itself does not claim adoption numbers, revenue, or performance measurements against Semrush or Ahrefs. Nothing was independently benchmarked here.

For a small SEO consultancy or an in-house team that is comfortable running Docker, the trade is straightforward: buy the compute, hold the data, and pay the vendors directly, in exchange for taking the operational work of running the stack. For a team without the operator, the paid suites remain simpler.

Source

• GitHub repository, primary source: https://github.com/SelmiAbderrahim/rankme.fast
• GitHub REST API for star and fork counts, queried 28 September 2026: https://api.github.com/repos/SelmiAbderrahim/rankme.fast
• RankMeFast homepage: https://rankme.fast/