What The Direct Index is, who runs it, and how it is paid for

The Direct Index measures whether AI assistants can find, read and act on a business's own website, and where they send people instead. It is independent, its method is public, and a named person is behind it.

What it does

It scans any business website. The scanner visits a site the way an AI agent does and answers four questions: can an agent reach it, can it read what the business offers, can it book or order there, and do assistants send people to it. The result is a score from 0 to 100 with every finding explained and a fix for each, in English and the languages of the business's country. It is free and takes under a minute.

It keeps score city by city. A city index takes every business of a kind that has its own website, from OpenStreetMap, runs the same scan on each and publishes the result: leaderboards, the share of sites that turn crawlers away, the booking and ordering vendors in use, the website builders, and how all of it changes from month to month. Cities can be compared with each other, ranked, and read as open data.

It observes the assistants. We ask ChatGPT, Claude and Gemini the questions people ask them, and record which businesses they name and where they send people: to the business's own site, to a marketplace, to a directory. Restaurant cities are asked every month, in the city's language in alternate months; the other kinds of business every quarter. Asked again on that schedule, this becomes the record of where agent traffic goes, kept in public.

It serves the direct path. For every business in the index we keep the record an agent needs to go direct: the booking link, the ordering link, the menu, the hours, the phone, which vendor sits behind each and whether that vendor's terms let an agent use it. Owners can correct the record through their claim; a browser verifies corrected booking and ordering links. Agents read it through the API, the MCP server, the A2A agent and daily feeds per city.

Where it stands

Counted on 15 October 2026 from the index's own statistics, which are refreshed hourly; dead domains are not counted as scanned. The scans started on 13 October 2026; the assistants' questions reach every index on a schedule, so the third number grows index by index. Every page states the sample and the date behind its figures, and a score is marked provisional until all four of its parts are measured.

Who runs it

The Direct Index is built and run by Antonio Curado, an engineer based in Brussels, Belgium. It started on 12 October 2026 with one question: when an assistant books a table, does it book with the restaurant or with a middleman, and what decides that? The answer turned out to be measurable, so the measurement is public, and any city can join it. One person runs the project. The site and the scanner run on Cloudflare's network; the methodology describes every number on the site, with its version.

How it is paid for

The project is funded by its founder. It takes no outside investment, no money from marketplaces and no advertising. If that changes, this page will say so first.

Reading is free: every report, every city page, the comparisons, the rankings, the search, the open data, the API, the MCP server and the A2A agent, with no account. Four things are sold. The deep report and the owner plan pay for the questions they ask the assistants; the agency plan and vendor verification pay for the work they add:

• The deep report, 19 EUR once: ten questions to each assistant about one business, their full answers, the nearest similar places, a fix plan and a PDF.
• The owner plan, 9 EUR a month: the assistants asked again every month with the differences, weekly rescans, alerts, the menu as a page and a monthly PDF.
• The agency plan, 79 EUR or 149 EUR a month: a client list, bulk scans, white-label reports, a lead finder and an API key with a higher scan limit.
• Vendor verification, invoiced per vendor: the action test on every site using a booking or ordering vendor, and a "works with agents" mark once it passes, stating the test and the date.

The score is computed the same way for everyone. What a plan buys is more frequent measurement, weekly scans and the assistants asked every month, which can move a score from provisional to measured sooner. Nothing paid changes how points are counted, a rank or a listing, and no placement is for sale.

Where the data comes from

• Business listings come from OpenStreetMap contributors, under the ODbL licence. The index is published under the same licence.
• Scan results come from our own visits: up to fourteen page requests per site plus a dozen small requests for well-known files, about thirty in all, each carrying signed headers that name directindex.org so any site can verify the visit. A robots.txt rule for DirectIndexBot stops the visit; the deeper pages stay within robots.txt; the home page is fetched as any visitor's page view is. A real browser takes a second look when the plain fetch is turned away, and runs the action test, which stops before the confirmation step. We never submit a form, place an order or make a booking.
• The assistants' answers come from the public APIs of OpenAI, Anthropic and Google, with web search on, asked the way people ask. Each answer is stored with the model that gave it and the date.
• Vendors' terms are read by a person, and each verdict carries the page it was read from and the date. A vendor who finds a reading wrong can write to us and it is corrected the same day.
• A model helps in five places: it classifies the destination hosts in the assistants' answers once each, into own site, vendor, marketplace, directory or press; it writes one dated sentence per business and round on why an assistant sent people elsewhere, from the stored answers and the scan's facts; it writes four sentences of commentary per monthly snapshot from the page's numbers, checked number by number; it reads a menu PDF into a page; and it answers questions on the ask page from the index's own tools. None of this touches a score, and a wrong judgement can be reported by mail.
• Nothing is bought from data brokers. The personal data we keep is the email address a person gives us to claim, correct or buy, and the contact details OpenStreetMap contributors recorded for a business; the privacy page says what happens to it.

If a page is about your business

Every business page has a free rescan and a claim form. A rescan settles most disputes, and the dispute button on every report opens the contact form below with the business filled in. A claim lets you correct the booking and ordering links, the menu and the hours yourself, through a signed link in the mail and with no account; a browser verifies corrected booking and ordering links before the record calls them verified. Language on every page stays factual: a site is "not reachable by agents", never "bad". If you own a site and want its page removed, ask and it will be removed.

Our own site

This site publishes what its own readiness check looks for: a sitemap, crawler rules, Content Signals, Markdown for every page, an API catalog, an MCP server, an A2A agent, published skills and signed requests; it has no accounts, so it publishes no OAuth. Cloudflare's independent checker rates directindex.org Level 5, Agent-Native. The ask the index page answers questions with the same tools the MCP server gives to agents, so what agents can do here is what we use ourselves.

Principles

1. Independent

   No money from marketplaces. No placement for sale. Scores are never for sale.

2. Open

   The methodology is public and versioned. Data is published under an open licence.

3. Honest

   Provisional scores are labelled. Samples and dates are stated. Every page has a free rescan and a dispute form.

4. Practised

   Our own site passes the standards it checks for, and our crawler can be verified by any site it visits.

Contact