Not the way it works on WordPress. Framer hosting has no server-side hook for your code, so the agent cannot install directly on the site. The free structural checks and tools still read every page from outside, and the route the agent uses on other hosted builders, a Cloudflare Worker in front of the site, applies here too once the domain's DNS sits on Cloudflare.
Because Framer gives you no server to put it on. On WordPress, the agent is a plugin that runs inside WordPress's own PHP layer; on a custom-coded site, it is one file in the middleware or a CDN worker. Framer hosting runs no code of yours at request time, so there is no layer to drop a plugin or a file into.
That is a limit of the hosting model, not a decision Reflexa makes about the site. The same is true of other fully hosted builders, Shopify, Webflow and Squarespace among them; the fix described on what the agent does on your site is the same one, just applied through a different route, where a route exists at all.
The structural checks and the five free tools, in full. They crawl the site's public pages from outside the way an AI engine would, so they need no install, no dashboard access and no account changes to run. Open the free check at dashboard.reflexatechnologies.com/audit against the live domain and every finding comes back the same as it would for any other site; see the 23 structural checks, one by one for what each one looks for.
| What it serves | What it needs |
|---|---|
| Structural checks (crawler access, schema, entity clarity, JS-only text) | Nothing; runs from outside on any public page |
| The five free tools | Nothing; same, no signup |
| Counting AI visits page by page | A place to run: a Cloudflare Worker, available if the domain's DNS sits on Cloudflare |
| The four switches applied automatically | Same: a place to run, which the Worker route provides where it is set up |
Two things, both inside Framer's own site settings. Framer lets you override robots.txt directly, so the AI-crawler groups for GPTBot, ClaudeBot and PerplexityBot can be added by hand; the crawler-access check shows exactly which ones are missing. Framer also offers a custom head code field, which injects a snippet into every page's head; that is where Organization structured data can be placed, and the schema and entity check shows what the page currently says, if anything.
Neither of these needs the agent installed. They are manual edits a site owner makes once inside Framer's own tools, checked against the free tool's output rather than guessed at.
Yes, under one condition: the domain's DNS has to be pointed through
Cloudflare with proxying turned on, the same requirement that
unlocks the full agent on Shopify, Webflow and Squarespace. Where
that condition is already met, a Cloudflare Worker in front of the
site works the same way it does on any other closed host: it sees
the whole response, so it can count every AI visit and apply the
four switches (crawler access, Organization data, the sitemap, and
/llms.txt) without anyone touching the dashboard by
hand. See the Cloudflare
Workers install guide for the one file it takes.
Where the domain's DNS does not sit on Cloudflare, Framer stays at checks and the two manual fixes above; it cannot count visits or flip a switch on a host it has no way to run on, until that route is in place.
The difference is entirely about where the agent runs, not what it does once it is running.
| WordPress | Framer | |
|---|---|---|
| Who installs it | You, no developer | No one, directly; a Worker once DNS is on Cloudflare |
| Where it runs | The Reflexa plugin, inside WordPress | Nowhere by default; a Cloudflare Worker in front of the site |
| Requirement | None beyond the plugin and a token | Domain's DNS on Cloudflare, proxied |
| If the requirement isn't met | N/A | Checks and manual site-settings fixes only |
Why it matters: a Framer site is never locked out of the free checks, only out of the parts of the agent that need a place to run; moving the domain's DNS to Cloudflare is the one step that closes that gap.
Bottom line: Framer needs no developer for the free checks and two manual fixes; the full agent needs a Cloudflare Worker in front of the site, which is a DNS change away rather than out of reach, unlike on hosts where the proxy route rarely applies at all.
Framer hosting has no server-side hook for your code, so there is no layer to install a plugin or a file into. WordPress has that layer; Framer does not, which is the whole difference.
The structural checks and the five free tools. They crawl the site's public pages from outside, the same way an AI engine would, and need no install and no account changes to run.
Two things, both in Framer's own site settings: override robots.txt, and add custom head code, which is where Organization structured data can be placed by hand. The free checks show exactly what is missing in each.
Yes, under one condition: the domain's DNS has to be pointed through Cloudflare with proxying turned on, the same requirement that unlocks the full agent on Shopify, Webflow and Squarespace. Where that is set up, a Cloudflare Worker in front of the site can count AI visits and apply the four switches without anyone touching Framer's settings by hand.
Yes. It belongs to this one site, lives in an environment variable rather than a committed file, and the domain proof refuses it from any other host, even another site on the same Reflexa account.
The free check reads your site the way AI crawlers do and shows every finding with its evidence, no card and no install needed.