launchdog
Sign in Connect

/guides/get-named-by-chatgpt-and-claude · checked October 29, 2026

How to get your product named by ChatGPT and Claude

When someone asks an AI assistant where to launch, what to use or what’s best, it names a handful of products and cites a few pages. We wanted to know how that works for a new product, so we asked ChatGPT, Claude, Gemini and Perplexity the questions our own users ask — 40 answers — and wrote down who they named and what they read. Here’s what we found, and what we’re doing about it.

What we asked

Ten questions a maker might type: eight that don’t name us (“What are the best Product Hunt alternatives in 2026?”, “Is there a product launch platform that works over MCP?”, “What’s a good weekly product launch board for indie makers?”) and two that do (“What is Launchdog?”). Each went to ChatGPT, Claude, Gemini and Perplexity with web search on, on October 28, 2026, the day after our own guide pages went live.

What we found

  • Asked about us by name, all four got us right: eight of eight answers described Launchdog correctly.
  • Asked the open questions, they almost never named us: one answer in 32. A new product isn’t on the pages the assistants read yet.
  • They learned about us from other people’s pages. Three of the four assistants learned what we are from someone else’s page — a SaaS directory listing, an MCP server directory, other launch sites — not our own. Only 3 of the 40 answers cited launchdog.dev at all.
  • They read lists and directories. The most-cited source across all 40 answers was a competing launch board’s own blog: its guides were cited in 16 answers. Then GitHub (8), SaaSCity (6), Reddit (6) and other boards’ “best Product Hunt alternatives” posts (4–5 each).
  • Names get mixed up. One assistant described a different company with a similar name; another recommended a similarly named site. A short, unique name and consistent descriptions everywhere help.

What that means for a new product

Assistants with web search behave like a careful reader with ten minutes: they search, open the top few pages, and repeat what those pages agree on. So the work is mostly being on the pages they open, and making your own pages easy to quote when they do.

1. Get onto the lists they cite

  • Ask the questions your buyers ask, with web search on, and note the pages cited. Those pages are your target list.
  • Pitch the authors of the lists and comparisons honestly: what your product is, in the shape of their table, and where it fits. We pitched four “Product Hunt alternatives” articles this way.
  • Get listed in the directories of your niche (for us, MCP server directories and launch-board lists). Keep each listing accurate: assistants quote them over your own site.

2. Publish things worth citing

  • Guides that answer the question directly, dated and specific. The board that wins citations in our category does it with its own guides.
  • Original numbers, like this page. Other writers and assistants cite data they can’t find elsewhere.
  • Fair comparisons with the alternatives, including when they’re the better choice. A page that only praises you reads as an ad and is quoted as one.

3. Make your own site easy to read

  • Say what it is in one sentence, in visible text near the top of the home page, the same sentence everywhere (your meta description, your directory listings, your social bios).
  • Let the search crawlers in. Check robots.txt for OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, PerplexityBot and Googlebot, and check that your CDN’s bot protection doesn’t block them either.
  • Add an llms.txt: a short Markdown summary with links to your key pages. It’s optional, but it’s ten minutes of work.
  • Structured data (Organization, SoftwareApplication, FAQ) and a clean share card help every reader, human or not.

The llms.txt and AI crawler checker checks the last three in one go.

4. Measure it the same way every month

Answers vary from run to run, so don’t read much into one. Keep a fixed set of prompts, run them monthly with the same assistants and settings, and count three things: how often you’re named without being asked about, how often your own site is cited, and who gets cited instead. We’ll publish our next runs here as the numbers move.

Where Launchdog comes in

A launch is one of the few moments a new product gets written about, listed and linked in many places at once — the pages assistants read. Launchdog is a weekly launch board you launch from your AI coding agent: your launch gets a permanent listing page with a followed link to your site. Launch in a few places (see where to launch) and the lists start to include you.

Questions

How do I get ChatGPT to recommend my product?

Assistants with web search answer from pages they find: mostly “best X” lists, directories, Reddit threads and comparison pages. Get your product onto the pages they already cite for your category, make your own site easy to read and quote, and let their search crawlers in.

Does llms.txt help AI assistants find my site?

It’s cheap and does no harm, but no major assistant has said it reads llms.txt when answering. What decides whether an assistant can read your site at all is robots.txt (and any firewall in front of it).

Which crawlers do I need to allow?

For answers and citations: OAI-SearchBot and ChatGPT-User (OpenAI), Claude-SearchBot and Claude-User (Anthropic), PerplexityBot, and Googlebot. GPTBot, ClaudeBot and Google-Extended collect training data; blocking them is a separate choice.

How do I know whether assistants mention my product?

Ask them, the same questions, on a schedule. Write down a fixed set of prompts your buyers would type, run them through each assistant with web search on, and record who is named and which pages are cited. One run is a sample; the trend over months is the signal.