AI Search & GEO

Pwodui llms.txt

Describe your site, add the pages you most want AI crawlers to read, and get a correctly-formatted llms.txt file. It is the emerging markdown standard you serve at the root of your domain to hand language models a clean, curated summary.

✓ Rezoud Markdown ki valab Ekri nan navigatè ou Pa gen enskripsyon
Rendered as a blockquote directly under the site name. Keep it to a single sentence.
Llms.txt ki te kreye

Kijan pou bati yon llms.txt dosye

Kat etap soti nan yon paj blan nan yon dosye curated AI crawlers ka li.

1

Dekri sit ou a

Enskri non sit ou a, URL paj lakay ou, yon rezime yon liy ak yon paragraf kout sou sa a pou yon AI crawler konnen ki sa sit ou a se.

2

Ajoute paj ou yo

Groupe paj ou vle pi plis ki citées nan seksyon tankou Dokiman, pwodwi oswa Blog, chak ak yon tit, yon URL absoli ak yon notifikasyon optionnel.

3

Tcheke prèv gade an vityèl

Markdown la mete ajou pandan w ap ekri. Avètisman flage rezime ki manke yo ak nenpòt URL ki pa yon lyen https absoli.

4

Sèvi ak li nan /llms.txt

Copy the file and publish it at the root of your domain (https://yourdomain/llms.txt), served as plain text or markdown.

llms.txt questions

llms.txt se yon pwopozisyon konvansyon: yon sèl markdown dosye plase nan ranje a nan domèn ou ki di gwo lang modèl sa sit ou a sou ak pwen yo nan paj yo ou pi renmen yo li. Li kòmanse ak yon H1 non sit, yon blokaj yon liy blokaj rezime ak yon dekriman kout, Lè sa a, lis lyen kuratè ki klase anba H2 seksyon tit.
At the root of your domain, reachable at https://yourdomain/llms.txt, served as plain text or markdown (text/markdown or text/plain). That fixed, predictable path is the whole point, since crawlers look there the same way they look for /robots.txt. The convention also allows an optional /llms-full.txt containing the expanded full text of your key pages inline, for models that want the content without a second fetch.
Be honest with yourself here: it might, but there is no guarantee. llms.txt is a proposed convention with limited and uncertain adoption. Most major AI crawlers do not yet commit to reading it, and publishing one does not make ChatGPT, Perplexity or Gemini cite you. Citations still depend on being indexed, being genuinely relevant and being trustworthy. Treat llms.txt as a cheap, tidy way to hand a well-structured summary to any model that does look, not as a growth hack or a ranking lever. That is the same honest stance we take on all AI-search visibility.
They solve different problems. robots.txt controls crawling: which paths a bot may or may not fetch. An XML sitemap is a machine-readable inventory of every URL you want discovered, for search-engine indexing. llms.txt is neither a gate nor an exhaustive list: it is a short, human-readable, curated introduction written for language models, highlighting only the handful of pages that best explain your site. Keep all three, since they complement each other rather than replace one another.

Ou vle sa a kreye ak kenbe fre pou ou?

This tool runs entirely in your browser, so there is nothing to call and nothing leaves the page. On the managed service we build your llms.txt from your real site structure and regenerate it as pages change, alongside the rest of your AI-search work.

✓ Gratis zouti: 100% bò kliyan an Managed llms.txt: plan ki peye

Auto-konstrui ak kenbe yon llms.txt curated soti nan sitemap ou ak paj tèt ou, epi swiv si motè AI reyèlman site ou. Gade plan ak quotes sou paj la pri.

Gade plan yo →
Developer API

Itilize li pa pwogram

Every tool is available through one authenticated REST API on the Pro and API plans. Wire it into your CI, CMS, or dashboards.

curl -X POST https://seo.bike/api/v1/llms-txt \ -H "Authorization: Bearer sk_live_yourkey" \ -H "Content-Type: application/json" \ -d '{"site": "example.com", "pages": ["/docs", "/pricing"]}'
OK
{ "ok": true, "result": { ... } }
Jwenn yon kle API → Metre, pou chak apèl pri. Grav nivo pou zouti entèaktif anwo a.
Check every page of your site, not one at a time
Pro connects Google Search Console, fetches every page Google skipped, every broken page and every page that ranks low, and writes what to fix on each one.
See how Pro works