A new year is a useful time to ask an old question: how should people find good products on the web?
Mydentify first answered that question in 2004. It was a web directory built around submitted links, browsable categories, and sponsorships. Thousands of submissions later, the lesson was not that directories were obsolete. The lesson was that a directory is only useful when someone does the work of organizing, checking, and explaining what belongs there.
The web did not stand still. Search engines became the main route to websites, and their algorithms changed what earned visibility. A useful category and a list of links were no longer enough. Pages needed stronger relevance, trusted references, original information, and a good experience after the click. Mydentify had to move with that change instead of protecting a format simply because it had worked before.
Traffic patterns changed too. Discovery spread across search results, app stores, software marketplaces, social feeds, newsletters, communities, and launch platforms. More recently, AI assistants began answering questions directly and choosing which sources to retrieve or cite. People still needed help finding products, but they no longer arrived through one predictable front door.
Going with the flow was necessary. It also made the original lesson clearer. Search is good at retrieving pages. Feeds are good at showing what is new. Neither always gives someone a stable, well-checked way to compare products by the job they need to do.
In 2026, Mydentify is returning to that foundation. This is not a restoration of a link page from the early web. It is a new kind of directory: one designed around what a person wants to accomplish, supported by structured evidence, and readable by both humans and software agents.
What disappeared when directories faded#
Search engines made the open web easier to query, but they also changed what discovery feels like. A search result can retrieve pages that mention the right words without helping a person decide which product fits a job. Marketplace categories often reflect how vendors describe themselves. Social feeds reward what is current. Review sites can be useful, but they usually cover only products that fit their commercial model.
A well-run directory does something different. It creates a stable map:
- Products are grouped using consistent language.
- Listings remain available after a launch-day spike ends.
- A reader can compare like with like.
- The publisher can show why a record is included and when it was checked.
- Narrow products can be discovered even when they do not dominate search results.
That model is useful again because the web is becoming harder—not easier—to evaluate. There are more products, more generated pages, and more ways for a weak claim to look polished.
What Mydentify is rebuilding#
The rebuilt Mydentify has two connected directories.
The first organizes products by intent: the outcome someone wants, such as finding initial users, improving a landing page, or understanding how a product appears to AI systems. Product matches should be explainable. Sponsorship is labeled and does not silently alter independent rankings.
The second is the startup directory dataset: a researched collection of places where founders can list or promote a product. It lets a founder compare category, pricing, link type, Domain Rating, and Mydentify's own Directory Score. Domain Rating is useful context, but it is not allowed to stand in for relevance, editorial quality, or likely return.
Those choices reflect the original directory instinct—organize the web for a person—but with a stronger data model underneath it.
A directory should explain its judgment#
Old directories often offered a category and a link. Modern directories can do better.
For a directory destination, Mydentify records the website, the submission path when known, the kind of product it accepts, whether a listing appears free or paid, the observed link type, Domain Rating, verification status, and a concise description based on real source material. Unknown values are omitted instead of presented as facts.
For product discovery, Mydentify separates claims, evidence, ownership, and sponsorship. That separation matters. A product owner should be able to correct a profile without purchasing influence. A sponsor should be able to buy a clearly marked placement without purchasing an organic ranking.
The result should be boring in the best sense: clear records, useful filters, durable URLs, and no mystery about what a label means.
Built for people first, readable by agents too#
People remain the primary audience. A page should make sense without special software: descriptive headings, visible evidence, normal links, and plain explanations.
The same discipline also helps crawlers and AI agents. Mydentify publishes server-rendered pages, canonical URLs, XML sitemaps, JSON-LD, public JSON feeds, Markdown representations, an llms.txt file, and an AI usage policy. These surfaces reduce ambiguity about what a page represents and where the canonical record lives.
They do not guarantee a ranking or a citation. Agent-readable publishing is an accessibility and interoperability practice, not a magic submission form for AI answers. The deeper explanation is in why Mydentify was rebuilt for people and AI agents.
New year, old conviction#
The conviction behind Mydentify has not changed: someone should make the web easier to browse.
What has changed is the standard. A useful directory in 2026 needs more than a large count. It needs a coherent taxonomy, distinct descriptions, working destinations, honest unknowns, visible commercial labels, and enough structure for other systems to interpret the collection correctly.
That is the work now. Mydentify is being rebuilt as a practical discovery layer for founders, product buyers, search engines, and agents—without pretending those audiences all need the same interface.
Start with the directory dataset, browse directory categories, or read why web directories are making a comeback.
See How We Checked It
We name the source, record the review date, and separate reported claims from our own checks. Read the methodology and editorial policy before citing a finding or sending a correction.
Put this guide into practice
Open the ai discovery tools hub, then choose the check that matches your next decision.
Found an outdated fact? Report a correction.
Back to articles