TheServiceMarketing Guys

A free tool

Generate your Open Service Profile.

It also writes your llms.txt. Enter your website. The tool reads your own pages, and your Google listing if you choose it, and drafts the files an AI assistant reads to learn who you are, where you work, when you are open, and how to request a booking.

One rule governs every line: a fact your pages do not state becomes a TODO line, never a plausible default. No price, hour, service, or certification is invented. Ratings and review counts are never copied; the standard forbids them.

Find your place id

Facts from the listing fill only what your site lacks, and never override it. One lookup is charged to a shared daily budget. If the budget is exhausted, clear the listing and run again.

Your booking policy

Please confirm these booking rules. The tool does not extract them from your pages. Answer what you know. A question you leave blank stays a TODO line in the draft and holds back the Level 1 claim until you answer it.

In whole days from today. 0 is today, 1 is tomorrow, up to 60.

Weekend requests
Do you take urgent calls outside your published hours?
How do you confirm a request?

The three standard request windows (Morning, Afternoon, Evening) are written by id only, with no clock times; the standard applies its own default boundaries in policy mode. Add your own times in the file if yours differ.

About your business

The tool derives these where one fact settles them (one recognized trade gives the service model; a state in one time zone gives the zone). Where the facts do not settle them, it asks you rather than guess.

Do you work at the customer's location, at your shop, or both?

Asked when your pages name no recognized trade, or more than one. Your answer always wins over the value the draft derived from your trade.

Asked when your state spans two zones. Your answer always wins over the zone the draft derived from your state.

Your robots.txt choices

The robots.txt section only writes directives you choose. Until you answer, it is a commented template that changes nothing.

May AI assistants read your public pages (search, answers, fetches a user asks for)?
May your pages be used for model training?

These are your choices, not the standard's. The tool never writes an Allow you did not state.

Up to twelve of your pages are read. We cache complete drafts for seven days and record usage for rate limits and paid lookups. We do not contact your business.

What the files are

  • open-service-profile.json goes at /.well-known/open-service-profile. It is the Open Service Profile manifest: your identity, locations and hours, services, coverage, and booking rules, in the shape the standard defines.
  • llms.txt and llms-full.txt are the plain-text guides assistants read first. The short one summarizes; the long one carries every fact with its source and says how each derived value was produced.
  • The robots.txt section names the AI crawlers and writes only the access you chose in the two robots.txt questions. Merge it into your existing robots.txt; until you answer, it is a commented template.
  • jsonld.html is the Level 1 JSON-LD for your home page: a LocalBusiness node with a ScheduleAction, built from the same facts. Paste it inside <head>.

The standard makes no claim about search results or where an assistant places you. These files organize business information for assistants to read. The interoperability results are still pending. Every booking is a request; you confirm. The standard itself: Open Service Profile.

From the command line

The same generator runs locally with Node 20 or newer and no dependencies. It writes the five files into a folder and prints the TODO list: npx open-service-profile init <url>

Open Service Profile generator | The Service Marketing Guys