How to add llms.txt to Framer (native support + fallbacks)

Framer ships built-in llms.txt support: published Framer sites can serve the file at yoursite.com/llms.txt automatically, generated from your pages. Check your site's settings panel for the toggle, verify the output, and use the edge-layer fallback only if your plan or site doesn't have the feature yet. Here's the whole flow.

Step 1: Check whether it's already on

Framer was early among site builders in adding AI-visibility features, llms.txt included. Two quick checks:

1. Visit yoursite.com/llms.txt on your published custom domain. Markdown = you're live already.
2. If it 404s, open your site settings in the Framer editor and look through the General/SEO panels for an llms.txt option to enable. Feature placement shifts as Framer iterates, so the settings panel — and Framer's own docs — are the source of truth for exactly where it lives on your plan.

Because platforms move faster than guides, treat this as the pattern: check native support first, build workarounds second. Framer is usually in the "native support" column.

Step 2: Understand what Framer generates

Framer builds the file from your site's structure — page names and descriptions become the file's links and annotations. Which means the generated file is only as good as your page metadata:

• A page called "Page 1" with no description contributes a junk line.
• "Pricing — Plans from $12/mo" with a real description contributes exactly what an AI assistant needs.

So the optimization path on Framer isn't editing the file — it's tightening the inputs: give every page a descriptive title and a real SEO description in its page settings. That work pays off in search snippets and in the generated llms.txt simultaneously.

Step 3: Validate the output

Auto-generated files are worth checking, not trusting. Run yoursite.com/llms.txt through the free llms.txt Validator — it verifies the H1/summary/link structure against the spec and flags malformed lines. Compare what Framer produced against good hand-written examples: if the generated file lists forty pages where a curated file would list twelve, that's the known trade-off of auto-generation — acceptable, but worth knowing.

Fallback: no native option on your site?

If your plan, region, or site configuration doesn't expose llms.txt yet, the universal workaround applies (same as Webflow):

Serve it from the edge. With your DNS on Cloudflare, a tiny Worker bound to the route yoursite.com/llms.txt returns your hand-written file as text/plain, while every other request passes through to Framer untouched:

export default {
  async fetch() {
    const body = `# Your Site\n\n> One line about what you do.\n\n## Key pages\n- [Pricing](https://yoursite.com/pricing): Plans and costs\n`;
    return new Response(body, {
      headers: { "content-type": "text/plain; charset=utf-8" },
    });
  },
};

Write the content itself with the llms.txt Generator — and the silver lining of the fallback is full control: you get a curated file rather than a generated one.

Step 4: Check the rest of the pipeline

llms.txt is the index; citations come from the pages it points to. Framer sites render server-side HTML (good — AI crawlers don't run JavaScript), but content structure still decides extractability. Run your homepage through the AI-Readiness Audit for the full checklist — crawler access, llms.txt, schema, and answer-first structure in one pass.

Bottom line

On Framer, llms.txt is mostly a check-and-verify task, not a build task: confirm the native file is live, feed it better page titles and descriptions, and validate the output. Only reach for the Cloudflare fallback if your site genuinely lacks the feature — and either way, spend the saved time on the content the file points to.