You probably arrived here from indexforai.net. That site is a machine-readable directory designed for AI agents. This page is for humans — the short version of what we do, why it matters, and where to go next.
A SaaS that helps businesses get found, cited, and recommended by AI agents — ChatGPT, Claude, Perplexity, Google AI Overviews, and Bing Copilot.
AI assistants increasingly read websites before answering business questions. The problem: most sites were built for humans and search engines, not for language models. Important facts about your business — what you do, where you operate, who you serve, what proof exists — are scattered across pages, buried in marketing copy, or expressed in ways an LLM has to guess at.
We scan your site, score how well AI agents can read it today, and generate a small set of machine-readable files that make your business legible in the format agents expect. The files are yours — you publish them on your own domain. We do the audit and the file generation.
30-second homepage scan or full-site crawl. We check structured data, AI-bot allowance, freshness signals, and how clearly your business identity reads.
llms.txt, llms-full.txt, ai-info.json, business-profile.json, plus four Schema.org JSON-LD blocks (Organization, WebSite, FAQPage, BreadcrumbList).
Weekly re-scan catches site changes and updates the files. Bot-analytics show which AI crawlers actually fetched your llms.txt and how often.
indexforai.net is the public, read-only directory of businesses that have run a Sofair scan and chose to publish their profile there.
AI agents and SEO crawlers browse the Index to find candidate businesses to cite when
answering questions like «best dental clinic in Tel Aviv» or «small SEO agency in Berlin».
Each business has a stable profile URL — indexforai.net/business/{slug}/ —
with the same set of files Sofair generates for your own domain. The Index has
no human UI by design; humans land on the page you're reading now.
Two domains, two roles: sofair.ai is the human side — accounts, billing, dashboard, this explainer. indexforai.net is the machine side — AI-bot endpoints, the directory, and bot-discoverable files. Both are operated by the same Sofair team.
llms.txt and ai-info.jsonEmerging conventions for telling AI agents about a website in a structured, predictable way.
llms.txt is a short markdown summary of your site — what it is,
what pages matter, in what order an AI should read them. The convention was proposed in 2024
by Jeremy Howard / Answer.AI and is now read by every major LLM-backed search system. Think
of it as robots.txt's younger sibling, but for content not crawl rules.
ai-info.json is Schema.org JSON-LD describing your
business — name, contact, services, area served, hours, proof, who you are. It
includes custom AI-friendly fields (aiSummary, bestFor,
notFor) that complement plain Schema.org. Sofair generates both for its
customers — and for itself: read our own
/llms.txt and
/ai-info.json.
# Sofair — AI Visibility Platform > SaaS that publishes machine-readable files > so AI agents can cite your business correctly. ## Docs - [What is Sofair](/humans-about-index) - [Sofair Index](https://indexforai.net/) ## Files we publish - /llms.txt - /ai-info.json - /ai-sitemap.xml
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "Sofair",
"url": "https://sofair.ai",
"description":
"AI Visibility Platform.",
"aiSummary":
"We audit sites and publish
8 AI-readable files.",
"bestFor": [
"small businesses",
"SEO agencies"
]
}
Same engine, different surfaces. Each one solves a piece of the AI-visibility problem.
Install once in Chrome. Scan any website in 30 seconds while browsing — your own, a client's, a competitor's. See the AI-readiness score, missing files, and one-click handoff to sofair.ai for the paid pack.
Learn about Sofair ScannerWordPress plugin that deploys + maintains the 8 files inside your WP site. No FTP, no manual upload — connect once with a token from your Sofair account, the plugin syncs new versions automatically when we re-scan.
Get the plugin (wp.org)Public read-only directory at indexforai.net. Bots browse /llms.txt + /sitemap.xml there to find candidate businesses to cite. Opt in from your Sofair account to publish your profile.
Quick answers to what people ask first when they land here.
Search engines rank pages by content plus signals (links, time-on-page, mobile-friendliness). AI assistants don't rank — they pick a single answer. They need machine-readable facts about your business to use you confidently.
SEO gets you on the results page. Sofair gets your business profile into the AI's answer. They're complementary; we don't replace your SEO team.
sitemap.xml and robots.txt?sitemap.xml tells crawlers which pages exist. robots.txt tells them which pages to skip. Neither describes what your business does.
/llms.txt and /ai-info.json fill that gap — they give AI a structured summary about your business that doesn't change unless you change it. You keep your existing sitemap.xml and robots.txt; we add the missing files alongside them.
llms.txt?As of 2026 the major LLM-backed search systems are testing or using these conventions: ChatGPT (OpenAI), Claude (Anthropic), Perplexity, Google AI Overviews, and Bing Copilot.
Adoption is fragmented and changing — Sofair monitors which agents fetched your files via the bot-analytics view in your account dashboard.
Yes. The free scan reads your homepage, checks for /llms.txt, /ai-info.json, Schema.org JSON-LD, and robots.txt allowance for AI bots, then gives you a 0–100 AI-readiness score. No card required.
Paid tiers start with a one-time pack of 8 ready-to-deploy files. The one-time pack is what most small businesses pick first.
Yes — llms.txt is a markdown convention and ai-info.json is plain Schema.org JSON-LD. The specs are public.
Sofair's job is the audit (what's missing, wrong, or mis-categorized), the per-page evidence map (so you know why each fact ended up where it did), and the maintenance loop (re-scan + auto-update when your site changes). DIY works for one site; the subscription works when you're managing many.
Only if you opt in. The Index is for businesses that ran a Sofair scan and chose to publish their profile. Opting in means your business is visible at indexforai.net/business/{your-slug}/ for AI crawlers to find.
If you only want the files on your own domain and not in the public directory, that's the default — no opt-in required, no profile published.
Free scan. 30 seconds. No card. You'll get a score and a list of what to fix — even if you never come back.
Scan my site free