Who Builds Hydrogen SEO? Meet Sheesh Labs (and What's Coming Next)
Hydrogen SEO is built by Sheesh Labs, a small Bengaluru studio. Its three principles, why it published a benchmark it lost, the 0.0.14 to 0.0.20 fixes, and RankWave AI.
In this article 7 sections
Hydrogen SEO is built by Sheesh Labs, a small software studio in Bengaluru, India, founded in 2026. The studio has two products: Hydrogen SEO, the free WordPress SEO plugin (Beta, version 0.0.20), its Pro add-on and the hydrogenseo.com platform, all of which are live; and RankWave AI, an AI-powered SEO and content-optimisation toolkit for WordPress that is still in build and taking a waitlist. Its WordPress.org contributor handle is sheeshlabs, which is the name you see on the plugin listing.
That is the whole answer. The rest of this post is about how the studio works, because the way Hydrogen SEO gets built explains a few things about the plugin that a feature list cannot: why we published a benchmark our own plugin lost, why every fix has a version number attached, and why the next product is not a surprise.
What is Sheesh Labs?
A small studio, deliberately. The tagline is "Filling the gap": Sheesh Labs builds practical, AI-powered software for the web, and it goes, in its own words, where the software is most used and least modern. WordPress qualifies on both counts. It powers an enormous share of the web, and much of its plugin ecosystem carries a decade of accumulated weight that nobody has gone back to remove.
The studio's about page states three principles. They are short enough to quote in full, and each one maps onto a decision you can see in the plugin.
| Principle | What it says | Where you see it in Hydrogen SEO |
|---|---|---|
| 01. Useful over impressive | "Version numbers, not adjectives" | Every claim on this site points at a measured number or a release number, not a superlative |
| 02. The parts nobody fixed | Go after the accumulated weight in software people already depend on | Zero front-end JavaScript, no custom database tables, +1 query, a 4.6 MB package |
| 03. Lab in public | "Builds, notes and readouts are the marketing; if something breaks, the fix ships with a number on it" | The benchmark that found our own canonical bug, published before the bug was fixed |
There is no team page with headcount, no funding announcement, no named advisers. We are not withholding them for effect; they are not part of the story. The studio is small enough that "who builds it" and "who answers your support ticket" are the same people.
Why did the studio publish a benchmark it lost?
Because principle 03 is not a slogan if you only apply it when the numbers flatter you.
In August 2026 we built a rig to compare seven WordPress SEO plugins on identical hardware: Yoast SEO, Rank Math, All in One SEO, SEOPress, The SEO Framework, Slim SEO and our own Hydrogen SEO, one at a time, database restored between runs, median of 13 measured requests. The full benchmark and the methodology are both on this site.
Hydrogen SEO did not win. The SEO Framework won or shared seven of sixteen categories and Slim SEO five. Our plugin placed third on PHP time and memory, second on queries and autoloaded options, won "most efficient schema" and tied on creating no database tables. That is a respectable result for a plugin a few months old, and it would have been easy to write the headline around the categories we won.
We could not, because the same run exposed a real bug. Hydrogen SEO 0.0.14 printed no canonical tag at all, on any page, and suppressed the one WordPress core would have printed. Bare WordPress with no SEO plugin handled canonicals better than our SEO plugin did. It was the single worst result any plugin posted in any category, and it was ours.
We published it anyway, with the bug still in the shipping version, then fixed it and re-ran the entire field on 8 September 2026 rather than patching our own row. The full account of that decision is on the Sheesh Labs blog: we benchmarked our own SEO plugin against six rivals, and it lost. The short version: a benchmark that only ever flatters its author is worth nothing, and we wanted ours to be worth something.
What has been fixed, version by version?
"The fix ships with a number on it" is the part of the third principle that is easiest to check, so here is the fix history from the first benchmark to today.
0.0.14 was the version in the first benchmark run. It printed no canonical tag, added +7 database queries, cost +15.4 ms of PHP time and shipped as a 5.5 MB package of 683 files.
0.0.15 fixed the canonical bug at the root rather than the symptom. The cause was that canonical output was opt-in per context handler and only the homepage had been wired up, so posts, pages, categories, tags and author archives all lost the tag silently. A design like that re-breaks itself, so the fix moved canonical output into one central place that covers every context, honours a per-page override, and points paged archives at themselves instead of page one. Search results and 404s are left without a canonical on purpose.
0.0.16 changed no code at all. It rewrote the plugin name, description and directory tags so the listing says what the plugin actually does, moved the Beta notice to where it is read, and marked compatibility with WordPress 7.1.
0.0.17 was a security release: the AI generation, embedding and site-wide bulk endpoints now require the plugin's administrator capability rather than an edit_posts class capability, so lower roles can no longer spend the site's AI credits or write SEO meta to posts they cannot edit; alt-text generation checks edit_post on each attachment; featured-image upload requires upload_files; a stored and reflected cross-site scripting issue in the document <title> was fixed by escaping at the output sink; automatic slug-change redirects are only recorded for users who can manage redirects; and a redirection rule with an empty match path is no longer treated as a site-wide catch-all.
0.0.18 was the performance release, and it is the version measured in the 8 September re-run. The plugin had been reading a dozen of its own settings options one at a time on every public page view, and on a site that had never saved a settings screen none of those rows existed, so each read cost a query that found nothing. Priming them in one query took seven extra queries per page view down to one on a fresh install and none on a configured site. The same release stopped packaging the pre-build TypeScript sources alongside the compiled output, taking the download from 5.5 MB and 683 files to 4.6 MB and 322. Measured: canonical present in every context, +1 query, +12.9 ms, 4.6 MB.
0.0.19 added the one-click import module under Tools → Import SEO Data, covering Rank Math and Rank Math PRO, Yoast SEO and Yoast SEO Premium, and SureRank and SureRank Business, with a dry-run preview, batched imports with pause and resume, undo from history and a WP-CLI command.
0.0.20, the current release, extended the importer to All in One SEO, The SEO Framework, Slim SEO and SEOPress, including their Pro editions. The full procedure is in switch from another SEO plugin.
Sheesh Labs keeps the bug-by-bug version of this list at Hydrogen SEO 0.0.14 to 0.0.20: every fix. Two of those fixes, the queries and the package size, were not trade-offs. They were waste, and we only found them because we ran our own product through the same rig as everyone else's.
What does the studio's approach mean for the plugin?
Three things you can verify rather than take on trust.
It stays out of your database. Hydrogen SEO creates no custom tables and adds 0.4 KB of autoloaded options against a 49.8 KB baseline. Deactivate it and 50.0 KB is left autoloading; the benchmark found plugins leaving 60.8 KB and 64.3 KB behind. Everything lives in standard WordPress options and post meta.
It stays off your front end. All SEO output is server-rendered markup with zero front-end JavaScript and no extra requests. In fairness, all seven benchmarked plugins passed that test; the difference is in wp-admin, where Hydrogen SEO's dashboard is 131.2 KB against a 129.8 KB bare-WordPress baseline, with heavy assets loading only on its own screens.
It works without an account. Every non-AI feature, from titles and meta to schema, sitemaps, robots.txt, the llms.txt editor, redirects, audits and the health score, works with no account and no key. A free hydrogenseo.com account only unlocks the AI features and Hydrogen AI, the agent; when credits run out only AI generation pauses. Plans on the pricing page differ in credits and sites, never in features.
The trade-off the studio accepts, and says so, is maturity. Hydrogen SEO is Beta software. The established plugins have years of production history and we have months. Test on staging first.
What is RankWave AI?
The studio's second product, and the reason this post has "what's coming next" in its title. RankWave AI is an AI-powered SEO and content-optimisation toolkit for WordPress. It is in build, not shipping, and the only thing you can do on its site today is join the waitlist. Its planned cloud service will run at api.rankwaveai.com.
We are not going to describe features that do not exist yet; that would break principle 01 on the first product it applies to. What we can say is what RankWave AI shares with Hydrogen SEO: the same studio, the same three principles, and the same habit of measuring before claiming. RankWave AI's own site has already put that to work by reviewing Hydrogen SEO the way a stranger would, which is the kind of thing sister products are supposed to be too polite to do. Sheesh Labs explains what it is building with RankWave AI on its own blog.
For a sense of how the two products relate, the studio's Hydrogen SEO page is the canonical statement of what the plugin is for, written from the studio's side rather than the product's.
How do I reach Sheesh Labs?
- Hydrogen SEO support: the support page on this site. For anything about the plugin, the platform or your account, start there.
- The studio directly: sheeshlabsai@gmail.com, or the contact details on sheeshlabsai.com.
- RankWave AI: the waitlist and contact details are at rankwaveai.com.
- Security reports: email the studio address above. 0.0.17 was a security release; reports are read and shipped with a version number on them.
If you want to judge the studio by its work rather than its words, the WordPress.org listing has the changelog, this site has the benchmark with our loss in it, and the installation guide takes about five minutes on a staging site. Version numbers, not adjectives.
Frequently asked questions
Who owns Hydrogen SEO?
Sheesh Labs, a small software studio in Bengaluru, India, founded in 2026. It builds Hydrogen SEO (the free WordPress plugin, the Pro plugin and the hydrogenseo.com platform) and is building RankWave AI. Its WordPress.org contributor handle is sheeshlabs.
Is Sheesh Labs the same company as RankWave AI?
Yes. RankWave AI is a Sheesh Labs product, currently in build and taking a waitlist at rankwaveai.com. Hydrogen SEO and RankWave AI are sister products from the same studio, sharing the same principles but with separate sites and separate plans.
Why did Sheesh Labs publish a benchmark that Hydrogen SEO lost?
Because the studio's third principle is to work in public and ship fixes with version numbers. The first benchmark found Hydrogen SEO 0.0.14 printed no canonical tag. It was published with the bug, fixed in 0.0.15, and the whole field was re-run on 8 September 2026 rather than patching one row.
Is Hydrogen SEO ready for production sites?
It is Beta software at version 0.0.20. The core output is server-rendered standard markup, it creates no database tables, and it has been measured against six established plugins on the same rig. As with any beta, test on staging first and report anything odd via support.
How do I contact the people who build Hydrogen SEO?
Use the support page on hydrogenseo.com for anything about the plugin or your account, or email the studio at sheeshlabsai@gmail.com. For RankWave AI, use the contact and waitlist details on rankwaveai.com.