NewFastlook now supports Google AI Overviews & Perplexity citations.Explore resources
Platform · AI Feed

Keep AI engines fresh on you.

Fastlook publishes your pages as fast, server-rendered HTML in your own theme, with JSON-LD, an llms.txt index, sitemap entries and a Markdown version for agents. On publish, it pings IndexNow and submits your sitemap to Google Search Console.

Every paid plan from Launch ($49/mo). Free includes a preview.

  • /llms.txt
  • /sitemap.xml
  • application/ld+json

Sample data A fictional bike shop. One URL, two representations.

What is AI Feed?

AI Feed is how Fastlook serves your pages to AI engines. Each page is fast server-rendered HTML in your own theme with JSON-LD, an llms.txt index entry and sitemap entry, and a Markdown version for agents. On publish, Fastlook pings IndexNow and submits your sitemap to Search Console.

GET /agent-ready

Can AI agents read my site today?

The free Agent-Ready Check runs 14 scored checks on your site across discoverability, content, bot access control, and API, auth, MCP and skills. No account needed.

Agent-Ready Check · tidewaterbikes.example Sample data
4 of 14 passed

Discoverability

  • robots.txt present/robots.txtPass
  • XML sitemap/sitemap.xmlPass
  • Link headers (RFC 8288)Link:Missing
  • DNS for AI Discovery_agents SVCBMissing

Content

  • Markdown for AgentsAccept: text/markdownPass

Bot access control

  • AI bot rules in robots.txtUser-agent: GPTBotPass
  • Content Signals in robots.txtContent-Signal:Missing

API, auth, MCP & skills

  • API Catalog (RFC 9727)/.well-known/api-catalogMissing
  • OAuth authorization server/.well-known/oauth-authorization-serverMissing
  • OAuth protected resource/.well-known/oauth-protected-resourceMissing
  • auth.md agent registration/auth.mdMissing
  • MCP Server Card/.well-known/mcp/server-card.jsonMissing
  • Agent Skills index/.well-known/agent-skills/index.jsonMissing
  • WebMCP (in-browser tools)navigator.modelContextMissing
Each result comes with what was found and how to fix it.Check your site →

/llms.txt

What is llms.txt, and do I need one?

A plain-text index at the root of a site that points language models at its most useful pages, the way a sitemap does for search crawlers. Fastlook keeps yours in step with the pages it publishes.

GET /llms.txt Sample data
  1. # Tidewater Bikes1
  2. > Commuter and cargo bikes, fitted and serviced in our workshop.2
  3. ## Guides3
  4. - [How to choose a commuter bike](https://tidewaterbikes.example/guides/commuter-bike): Frame size, bars and racks for daily riding.4
  5. - [Cargo bike or trailer?](https://tidewaterbikes.example/guides/cargo-bike-or-trailer): Loads, storage and hills compared.
  6. ## Services
  7. - [Bike fitting](https://tidewaterbikes.example/service/fit): What a fit check covers and how to book one.
  1. 1

    The site name, as a title.

  2. 2

    One line on what the business does.

  3. 3

    Sections group your published pages.

  4. 4

    Each link carries a one-line description an agent can read before it fetches the page.

Accept: text/markdown

What do agents get instead of HTML?

The same URL, asked a different way. A browser asks for HTML and gets your themed page. An agent that sends Accept: text/markdown gets the content as clean Markdown, on pages served from the Fastlook feed. Flip the request and see.

Request as
Sample
GET /guides/commuter-bikeHost: tidewaterbikes.exampleAccept: text/html200 content-type: text/html; charset=utf-8Browser view shown: text/html; charset=utf-8.
How to choose a commuter bike

Pick a bike for the route you ride every day, not for a race.

What frame size do I need?

Stand over the top tube with a few centimetres of clearance, then check reach on a test ride.

Flat bar or drop bar?

Flat bars give an upright view of traffic. Drop bars add hand positions for longer rides.

Tidewater Bikes

The sample page as a browser renders it.

Size of the two samples on this page, measured in your browser

HTML1,258 bytes
Markdown338 bytes

On this sample the Markdown is 27% of the HTML. Real pages vary: theirs carry scripts, styles and markup this sample leaves out.

application/ld+json

How do engines know what a page is about?

Every published page carries JSON-LD that names the business, the page's place in the site, the article and its questions, so a parser doesn't have to guess from layout.

Rendered page Sample data

1Tidewater Bikes

2Guides / Commuter bikes

3How to choose a commuter bike

4What frame size do I need?

Stand over the top tube with a little clearance…

<script type="application/ld+json">
  1. {
  2. "@context": "https://schema.org",
  3. "@graph": [
  4. { "@type": "Organization", "name": "Tidewater Bikes", "url": "https://tidewaterbikes.example" },1
  5. { "@type": "BreadcrumbList", "itemListElement": [ "Guides", "Commuter bikes" ] },2
  6. { "@type": "Article", "headline": "How to choose a commuter bike",3
  7. "dateModified": "2026-09-28" },
  8. { "@type": "FAQPage", "mainEntity": [4
  9. { "@type": "Question", "name": "What frame size do I need?",
  10. "acceptedAnswer": { "@type": "Answer", "text": "Stand over the top tube…" } } ] }
  11. ]
  12. }

POST /indexnow

How do engines hear about a new page?

Fastlook tells them when you publish, instead of waiting to be found. This is a nudge, not a guarantee: each engine still decides when to crawl and what to index.

  1. 01Page published/guides/commuter-bike · server-rendered HTML in your theme200 OK
  2. 02Sitemap entry added/sitemap.xml · new <url> for the page/sitemap.xml
  3. 03llms.txt index updated/llms.txt · the page listed with its description/llms.txt
  4. 04IndexNow ping sentthe new URL submitted to engines that support IndexNowIndexNow
  5. 05Sitemap submitted to GoogleSearch Console told the sitemap changedSearch Console

/.well-known/

How do agents discover what my site can do?

A newer set of files at fixed addresses on your domain. They live on your own site, and the free Agent-Ready Check tests for each one and explains how to add any that are missing.

  • /.well-known/api-catalog

    Lists the APIs your site offers, in a format agents can parse.

  • /.well-known/mcp/server-card.json

    Describes an MCP server an agent can connect to.

  • /.well-known/agent-skills/index.json

    An index of skills your site exposes to agents.

  • /.well-known/oauth-authorization-server

    Tells agents how to authenticate against your APIs.

  • /.well-known/oauth-protected-resource

    Describes which resources need a token, and from where.

  • /auth.md

    Plain-language agent registration instructions, served as Markdown.

What your site serves today is one scan away.Run the free check →

<html data-theme>

Will the pages look like my site?

Yes. Pages render in your own theme: your colours, type and spacing, read from your site. Not a Fastlook template with your logo on top.

A bike workshop owner smiling at his phone among bicycles hung on the wall
People get your themed page. Agents get the structured feed.
tidewaterbikes.example

Tidewater Bikes

How to choose a commuter bike

Pick a bike for the route you ride every day, not for a race.

Book a fit
Theme A · serif headings, square corners
tidewaterbikes.example

Tidewater Bikes

How to choose a commuter bike

Pick a bike for the route you ride every day, not for a race.

Book a fit
Theme B · inverted surface, rounded

Illustrative Same content, two sample themes.

User-Agent:

Which AI bots actually came?

Fastlook classifies AI crawler hits on each page, including GPTBot, PerplexityBot, ClaudeBot and Google-Extended, so you can see which bots visited and when.

Crawler log · tidewaterbikes.example Sample data
  • GPTBotGET /guides/commuter-bikeAI crawler
  • PerplexityBotGET /guides/cargo-bike-or-trailerAI crawler
  • ClaudeBotGET /llms.txtAI crawler
  • PerplexityBotGET /service/fitAI crawler
  • GPTBotGET /sitemap.xmlAI crawler
Analytics shows who crawled, ranked and mentioned each page.See Analytics →

Honest limits

What doesn't the AI Feed do?

Two things worth knowing before you rely on it.

Accept: text/markdownMarkdown is served on feed pages

Pages served from the Fastlook feed answer Markdown requests. If you publish into your own CMS, that CMS serves the page; run the Agent-Ready Check to see what it returns.

User-Agent:Bots are identified by what they declare

Crawler hits are classified from each request's user agent. Being crawled is a good sign, but it is not the same as being cited in an answer.

FAQ

Questions about AI Feed

What is AI Feed?

AI Feed is how Fastlook serves your pages to AI engines. Each page is fast server-rendered HTML in your own theme with JSON-LD, an llms.txt index entry and sitemap entry, and a Markdown version for agents. On publish, Fastlook pings IndexNow and submits your sitemap to Search Console.

What is llms.txt?

A plain-text file at the root of a site that lists its most useful pages for large language models, similar to how a sitemap lists pages for search crawlers. Fastlook keeps yours updated as pages are published.

Why does a machine-readable surface matter for AI search?

Answer engines quote what they can parse reliably. Pages served as fast HTML with clean headings and JSON-LD are easier to extract than pages that need JavaScript to render.

Does the AI Feed replace my website?

No. It adds structured, answer-shaped pages alongside your site, rendered in your brand's theme.

How do I know AI crawlers are reading it?

Fastlook classifies crawler hits on each page, so you can see which bots visited and when.

Start with AI Feed

One URL. A page for people, a feed for agents.

llms.txt, sitemap, JSON-LD and Markdown, kept in step with every page you publish.

  • ChatGPT
  • Perplexity
  • Gemini
  • Claude
  • Copilot
  • AI Overviews

Be the answer AI finds. Your AEO growth partner.

A founder in a black turtleneck, editorial black and white portrait with a warm orange glow