biro . works

BUILT FOR THE AI WEB

Sites agents can read, recommend, and trust.

Most websites still talk to humans only. Yours shouldn't. Biro ships sites that ChatGPT, Claude, Perplexity, and Google AI can find, parse, and cite, by default.

Get a quote See how it works

Structured data · llms.txt · semantic HTML · no JavaScript-only content.

§ 01 · WHY NOW

AI is the new top of the funnel.

  1. 01

    Prospects ask AI before they ask Google.

    ChatGPT browsing, Claude search, Perplexity, and Google AI Overviews now answer questions that used to start a search. The studios cited in those answers get the work.

  2. 02

    Structured data is what gets quoted.

    Schema.org markup with explicit prices, services, and contact info is what AI tools lift verbatim. Sites without it get paraphrased badly, or skipped entirely.

  3. 03

    JavaScript-only sites are invisible.

    Most AI crawlers don't execute scripts. If your content only renders on the client, the agent reads an empty page and moves on.

  4. 04

    The studios that ship this win the citation.

    Most sites still ignore agent readiness. Doing it well is a small technical investment with an outsized discovery payoff.

§ 02 · WHAT WE SHIP

Agent-readiness, by default.

Every site we build, host, or maintain ships with the following. No upcharge, no asterisk.

  • Schema.org markup

    Organization, Service, Offer, FAQPage, ContactPoint emitted as JSON-LD on the relevant pages. Validates against Schema.org and Google Rich Results.

  • llms.txt + llms-full.txt

    A short site summary at /llms.txt and a single-document full reference at /llms-full.txt for agents that prefer one fetch over a crawl.

  • Open robots.txt

    GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and CCBot all explicitly allowed. We can block any subset on request.

  • Server-rendered HTML

    Content lives in HTML, not JavaScript. Agents that don't execute scripts still read the full page.

  • Clean canonicals + hreflang

    One canonical per URL, honest cross-locale signals, no contradictions between sitemap and page meta.

  • Core Web Vitals

    Fast first paint so agents that render a viewport don't time out. Static delivery, sensible images, no client-side bloat.

  • Plain-prose copy

    Readable by humans first, summarisable by LLMs second. No interview-driven jargon, no AI-flavoured boilerplate.

  • OG, Twitter, sitemap

    The boring metadata, done right. Sitemap that only lists pages that exist. OG images that render in every preview surface.

§ 03 · PROOF

See it on this site.

If we ship it on our own site, we ship it on yours. Open the files. Read the source. Validate the schema.

NEXT

Ship a site agents can find.

Get a quote See plans & pricing