Loading…
Launch your AI tool
An AI tool has two audiences on launch day and only one of them is human. The other is the crawler assembling an answer to "best tool for X" three months from now, and it will decide whether your listing is quotable based on things most launch sites get wrong.
That is the part LetsLaunch is built around: every page has a plain-Markdown twin, the whole catalogue is published as a single text file, and the crawlers behind AI search are allowed by name rather than by default.
Concretely, and with the limits stated.
A listing models can actually parse
Append .md to your product URL and you get the page as plain Markdown, noindex with a canonical back to the HTML so it never competes with your listing in search. It exists purely so a model does not have to guess.
A place in the catalogue file
/llms.txt documents the convention and /llms-full.txt is the entire catalogue in one file. A model that fetches one URL gets every listing, yours included.
Structured data that will not break the page
All JSON-LD is serialised through one escaper, because product names and taglines are maker-supplied and an HTML parser closes a script tag at the first literal </script whatever the JSON says.
A followed link that is not for sale
Earned via the reciprocal badge, kept via the votes. Sponsored content on this site carries rel="sponsored" and never a followed link — which is the only reason the earned ones are worth anything.
Paste your URL
The form reads your site and prefills the name, tagline, description, categories and screenshots. You edit what it got wrong instead of typing ten fields from scratch.
Install the badge
Copy the embed from your listing's badge page and put it anywhere on your own site. Any real crawlable link back to your listing URL counts — it just must not be marked nofollow, ugc or sponsored.
The check publishes you
The first automatic check runs about an hour after you submit, and you can trigger a manual re-check yourself the moment the badge is live. When it passes, the listing publishes with a dofollow link and its launch clock starts from that moment — not from when you filled the form.
Collect 5 upvotes
You get 24 hours on the board to clear the bar. Clear it and the listing and its followed link are permanent; miss it and the listing is hidden from browse with one free relaunch available. We email a warning if you are short with time left.
The rule: a published listing has 24 hours from launch to collect 5 qualifying upvotes. Clear it and the listing and its followed link are permanent. Miss it and the listing is hidden from browse and its outbound link drops to nofollow — nothing is deleted, and you get one free relaunch. While LetsLaunch is still building up daily volume this rule is paused, so no listing is currently removed for missing the threshold.
Three tiers, all published. The free one is the standard path, not a trial.
Free Launch
Free
Everything you need to launch in public, forever free.
Premium
$19
Skip the badge upkeep and stand out on the board.
Premium+
$39
For teams who want a real content asset out of launch day.
Full breakdown, including ad placements, on the pricing page. No tier buys a followed link, a position on the board, or survival past the window.
Four surfaces, all indexed separately, all linking back to your product page.
The AI tools directory
A permanent, browsable view of every AI, automation and writing listing in the catalogue, ranked by votes.
Category boards
Each category is its own ranked page with its own URL, and a tool can sit in up to three of them.
Alternatives pages
The catalogue generates 'alternatives to X' pages, which is exactly the query shape people use when they are shopping for an AI tool.
The plain-text catalogue
llms-full.txt, regenerated as the catalogue changes, is the single file an answer engine is most likely to fetch whole.
This week's launches, ranked by upvotes — the board yours would join.
Nobody can sell you a citation in ChatGPT or Perplexity, and anyone offering one is guessing. What can be done is removing the reasons a model would skip your page — and most of those reasons are boring and technical rather than strategic.
The common failures: the page needs JavaScript to render its own content, the crawler is blocked by a default robots.txt nobody read, the description is marketing copy that never states plainly what the tool does, and there is no machine-readable version of the page anywhere. A listing here fixes the last three for you and the Markdown mirror fixes the first.
The part you control is the description. Write what the tool does in the first sentence, in the words someone would use to search for it, before any positioning. A model extracting an answer has no patience for a build-up, and neither does the reader it is answering.
Then check your own site: our free AI crawler check will tell you whether GPTBot and ClaudeBot can reach it at all, which is worth knowing before you spend a launch week driving models toward a page they cannot read.
Paste the URL, let the form fill itself in, install the badge, and go live.
