Seonaut Review

Overview

SEOnaut is an MIT-licensed SEO crawler written in Go, self-hosted with Docker and MySQL, and also offered as a hosted service with a free Lite tier (one project, 500 URLs per project) and a $9-per-month Growth plan (five projects, 10,000 URLs, recurring audits), billed through Stripe with one month free on annual plans; the site claims more than 3,000 professionals use it. The code defines 79 issue types across page-level and site-wide checks: broken links and redirects including loops, missing or duplicate meta tags, heading order, hreflang problems, image and alt-text audits for accessibility, content quality, canonical tags, orphan and dead-end pages, HTTPS checks, and TTFB and timeout detection. Documented crawl options include bypassing robots.txt, following nofollow links, crawling from sitemaps or subdomains, checking external links, basic-auth crawls, and WACZ archive creation. Results land in Apache ECharts dashboards with CSV, sitemap, and WACZ exports, and the crawler identifies itself as SEOnautBot/1.0. The honest limits shape who should adopt it: there is no JavaScript rendering, since no headless browser exists in the codebase, so client-rendered pages will audit badly; there is no API, because all routes are session-based web UI; there is no team or role model, with single-user projects in the schema; and no AI features appear anywhere in the product or its site. Maintenance is active but unhurried: 900-plus commits since 2022, the most recent in May 2026, and no tagged releases, so installs track the :latest container image. Source builds need Go 1.25 and the compose file runs MySQL 8.4, serving plain HTTP on port 9000 with signup rather than a default account. Against Screaming Frog the trade is direct: no license cost and repeatable self-hosted audits, in exchange for JS rendering and a commercial feature set.

How to install

1. Docker path from the README: git clone https://github.com/stjudewashere/seonaut.git, cd seonaut, then docker compose -f docker-compose.yml -f docker-compose.dev.yml up -d --build, and open http://localhost:9000.
2. Image-only install: docker pull ghcr.io/stjudewashere/seonaut:latest. From source without Docker: make run per docs/INSTALL.md.
3. There is no default account: you register through the /signup route on first run. The compose database credentials (MYSQL_ROOT_PASSWORD=root, database, user, and password all seonaut) are sample values to change before exposing anything.
4. Configuration lives in a config file at the repo root, overridable with SEONAUT_* environment variables; the crawler user agent is Mozilla/5.0 (compatible; SEOnautBot/1.0; +https://seonaut.org/bot), so exclude it with robots.txt if you audit sites you do not own.
5. The app serves plain HTTP on port 9000; the install docs describe fronting it with Nginx or Apache plus certbot for HTTPS.

Requirements

Docker and Git for the container path; Go 1.25 and Make to build from source, with MySQL 8.4 as the database in the compose file. A hosted option at app.seonaut.org removes the infrastructure work entirely.

Best for

Developers and consultants who want a repeatable, self-hosted technical crawler with no per-crawl or license cost, and whose target sites are server-rendered: agencies auditing client HTML, and site owners tracking broken links, redirects, metas, hreflang, and alt text on a schedule they control.

Not for

Auditing JavaScript-rendered applications (no headless browser in the codebase), teams wanting shared projects or roles (the schema is single-user), and anyone needing programmatic access (there is no API; exports are CSV, sitemap, and WACZ files from the UI).

Review notes

Assessed from the GitHub repository (including the issue-reporter source and route table), seonaut.org, and docs/INSTALL.md; we have not run a crawl. The code defines 79 issue types registered as 57 page-level and 15 site-wide reporters, which is the concrete answer to how much audit coverage the project ships.

Two corrections against our earlier record: the tool has no API (we previously marked api_available true, but every route is session-cookie web UI with no /api endpoints), and the widely repeated claim of 150-plus issue types does not appear in any primary source; the count in code is 79. We also listed integrations that do not exist; the only exports are CSV, sitemap, and WACZ.

Maintenance cadence is the honest risk flag: 534 commits in 2022, then roughly 110 to 130 per year, 33 in the trailing twelve months, and the most recent commit in May 2026. There are no tagged releases, so pinning is not part of the project's model and upgrades mean pulling the latest image. MIT licensing and commercial use are explicitly permitted per the support page.

The hosted tiers change the evaluation: Lite is free with one project and 500 URLs per project, Growth is $9 per month with five projects, 10,000 URLs, and recurring audits, which is the only place scheduled crawling is documented. Self-hosted, you bring your own scheduler via cron.

Verdict

A lean, honest, self-hosted technical crawler with 79 documented issue checks; no JS rendering, no API, and a slow-but-alive commit pace.

Similar Tools