About EffectHub

A browsable index of AI products and MCP servers, organised into categories and subcategories, with the things you actually decide on: what it does, what it costs, whether there is an API, who is behind it, what the alternatives are, and whether it still exists.

Entries are not published the moment they are imported. A record goes live only once it clears a completeness bar — enough verified fields and a description written for a reader, not scraped filler. Everything else stays in our import pool as a draft until it earns a page. That is why the catalog grows in waves instead of dumping a hundred thousand thin pages into the index.

Cards are written and maintained by EffectHub Research — our own editorial process, described below.

EffectHub is an independent catalog of AI tools at effecthub.net, live since August 2026. It is built and run by a small team that researches every listing from the vendor’s own pages, re-checks each tool’s website on a schedule, reads domain registration records, and publishes what the catalog counts as EffectHub Research. It is not the game-development platform at effecthub.com, and has no connection to it.

Every listing carries one number out of 100. It is derived at read time from what the catalog has measured about the tool — nothing is stored, so the figure can never drift away from the data it summarises, and nobody can edit it.

Four signals, weighted by how hard each one is to fake. A signal we have no data for is not scored zero: it is left out and the remaining weights carry the number, which is what keeps a tool published last week from being punished for not having a history yet. A trace of data counts as no data for the same reason — eleven page views are noise, and treating them as a verdict would rank a tool below an identical one nobody has visited.

• 35%Real demand — Outbound clicks to the tool, counted twice, plus page views on its listing. Attention we can observe rather than attention someone claims — weighed by how much of it there is, so a few dozen visits nudge the score and thousands carry the full share.
• 30%Reception — Visitor ratings, pulled toward the catalog average until enough of them land. Three glowing reviews move a score far less than thirty.
• 20%Corroboration — How many of our source catalogs list the tool independently. One directory is a listing; six is a track record.
• 15%Standing — Alive at the last check, and a listing complete enough to act on — priced, described, placed. A dead tool scores nothing here whatever else it has. A site that turns our checker away is not scored here at all: we could not look, so the signal is missing rather than zero.

A listing can be bought — see the pricing on the submission page. A rank, a score and a review cannot, and no advertiser has ever been shown this arithmetic before it ran.

The rule behind every card: its status and its Last verified date come from a check we run against the product itself — never from a claim on the vendor’s homepage.

The checker requests the tool’s own address and reads what comes back. A check fails when the domain no longer resolves, the server stops answering, the certificate is invalid, or the site answers 404, 410 or a server error. Being blocked by a bot wall is not a failure, and neither is a redirect to another domain. A parked or for-sale page answers like a live site, so the checker cannot catch one — an editor marks it by hand, and no later check undoes that. The methodology page lists exactly what the checker sees.

One bad response proves nothing: sites have outages, and hosts throttle bots. The first failure marks an entry possibly discontinued, and only a second failure in a row, on a later run, marks it discontinued. Re-checks run on a rolling schedule — the tools people open most are re-tested every three days, ordinary entries weekly, the long tail every 30 days. So on a long-tail entry a last-verified date a few weeks old is the schedule working, not a page we forgot.

• ActiveThe last check reached the product, or was kept out by a bot wall — a blocked request, not a sign the product is gone.
• Possibly discontinuedThe last check failed. The entry stays visible with the warning attached until the next check settles it.
• DiscontinuedTwo failed checks in a row, on separate runs — or an editor's call, such as an acquisition or a parked domain, which no later check reverses. The entry keeps its page and the reason we logged instead of being quietly erased.

Where this stands today. The checker covers the whole published catalog: every hour it re-tests up to 200 cards, never-checked ones first. A card starts with the status established during import research and carries the checker’s from its first visit; until then it reads Not yet verified rather than borrowing a date. When runs are missed the dates show it, and the cards claim no more than the dates support — we would rather describe the method and its actual coverage than imply the whole catalog was re-tested this morning.

Anything suspected dead drops out of rankings and recommendation lists immediately — it just keeps its page. When a tool is confirmed gone we do not silently delete the record: it keeps the date and the reason we logged, because “is this thing still alive?” is a real question and a blank page is not an answer.

Counters are literal. A number on this site is whatever the database returns — clicks, saves, ratings, tools in a category. We do not round up, pad, or display an aggregate we did not measure. If we have no data for a field, the field is empty instead of decorated.

No pay-to-rank. Money never moves a tool up an organic list, ranking, or set of alternatives. Anything commercial is a separate, visibly labelled slot, and when we sell a faster review it buys queue position, not catalog position. Outbound links that can earn us a commission carry sponsored/nofollow markup. If a listing is here because someone paid, you will be able to tell without reading a disclosure page.

Facts start at the source: the vendor’s own site — product pages, pricing page, docs, changelog — plus open, publicly available datasets and curated lists that tell us a tool exists in the first place.

From other catalogs we take facts only: that a product exists, its canonical URL, sometimes a category hint. We never reuse their copy. Every description on EffectHub is written by us from the vendor’s own material, which is also why an entry can sit unpublished for a while — a copied paragraph is faster, and worthless.

Fields we cannot confirm stay empty. We would rather show a short card than a confident wrong one, and pricing in particular is checked against the vendor’s page rather than inferred.

Founders and vendors can have their own entry corrected. So can you: if a fact here is wrong or a tool is missing, tell us and we will fix the record and re-verify it.

EffectHub is early and we would rather say so than imply a decade of history. The catalog is being built in curated waves, features land one at a time, and some sections are still empty. The verification method above is the standard we hold ourselves to — its automated half runs across the whole catalog, and when it falls behind the cards say so in their dates — which is why it is written on this page instead of buried in a footnote.

Corrections, removals, missing tools, deals, partnerships, press — one place: the contact page. Point at the page, tell us what is wrong, and we will re-check it.

Browse all categoriesBack to home