How to Submit Your AI Tool to Linkgo (And What the Curation Pipeline Actually Does)

You shipped something. Maybe it is an MCP server for Postgres, a small agent framework, or a hosted tool that does one thing well. The product works, the landing page is live, and now nobody can find it. Directories are one of the few places where people who are actively looking for a tool will actually land on yours, but most of them are a black box: you fill in a form and hope. This post explains how listing a tool on Linkgo works, and what the curation pipeline does behind the scenes, so you know what to expect.

Two ways a tool reaches Linkgo

Linkgo is a curated directory of AI tools, agents, services, MCPs, and models, and entries arrive by two routes.

The first is submission. The Linkgo site has a "Submit Your Tool" call to action for builders who want their product considered. If you built it, this is the direct path: you tell Linkgo it exists.

The second is discovery. A daily operator-run agent scans the places where new AI products usually surface first: Product Hunt, GitHub Trending, Futurepedia, There's An AI, and Futuretools. That means a tool can end up in the directory even if its builder never filled in anything, and it means a submitted tool is not the only way in. Submitting simply removes the dependence on being spotted.

What happens after a candidate shows up

Whichever route a tool takes, it passes through the same pipeline before a visitor sees it.

Pre-filtering. Not every candidate belongs in a curated directory. Obvious misfits are dropped early so that an operator is not reviewing noise.

Dedupe by normalized URL. The same product often appears under slightly different addresses: with and without a trailing slash, with tracking parameters, with a different capitalization. Linkgo normalizes the URL before comparing, so a tool that was already discovered through Product Hunt will not be entered a second time because it also trended on GitHub. If you submit something that is already in the catalog, expect it to be recognized as a duplicate rather than listed twice.

Category routing. Each entry is placed into exactly one of five categories: AI Agents, AI Tools, AI Services, MCPs, or AI Models. MCPs and agents are first-class categories rather than tags, so an MCP server is routed to MCPs, not buried under tools. Knowing which bucket you belong in before you submit is useful: it is the first question anyone browsing will ask about your product.

Enrichment. The entry is filled out with the things a browsing builder wants to compare at a glance: a screenshot, pricing information, and a description.

FAQ generation. Each entry gets an auto-generated FAQ section, so the page answers the questions people tend to ask about a tool before they have to click through to your site.

Operator approval. A person reviews the entry before it goes live. The catalog is curated, not a crawler dump, and this is the step that keeps it that way.

What to prepare before you submit

The pipeline can only enrich what it can find, so the best thing you can do is make your own product easy to read.

Make sure your landing page states plainly what the tool does and which kind of product it is: an agent, a tool, a service, an MCP server, or a model. If a person has to guess your category, the routing step has to guess too.

State your pricing on a public page. Pricing is one of the main ways builders filter a category, and an entry with no clear pricing is harder to shortlist.

Have a clean, current screenshot of the real interface. Directory visitors judge maintenance from screenshots, and a stale one works against you.

Use one canonical URL and stick to it. Because duplicates are detected by normalized URL, consistency avoids confusion between your homepage, a docs subdomain, and a launch page.

What submission does not guarantee

A submission is a candidate, not a placement. Because an operator approves every entry, a tool may be reviewed, routed to a different category than you expected, or not listed. Linkgo also makes no claim of exhaustive coverage, and it is not a rating engine: there are no aggregate scores or vote leaderboards, so a listing is an introduction rather than a ranking. Review timing is not something to count on; if your launch date is fixed, plan around the possibility that the entry appears after it.

Where to go next

If you have a tool, agent, or MCP server that fits one of the five categories, the path is short: tidy your landing page, pick your category, and use the "Submit Your Tool" option on Linkgo. If you are not sure whether a listing already exists, search the catalog first.

Linkgo is a curated directory of AI tools, agents, MCPs, and models — browse the catalog at linkgo.dev.