llms.txt
Control How AI Describes Your Business

llms.txt is an emerging open standard — a plain-text file hosted at the root of your website (like robots.txt) that gives AI language models structured, authoritative information about your business. Instead of LLMs like ChatGPT and Claude guessing who you are from scraped web content, they can read your own curated brief. The standard was proposed in late 2024 and is gaining rapid adoption as businesses realise that AI assistants are becoming primary discovery channels. The earlier you establish your llms.txt, the more training data cycles and crawl events your accurate information has to influence how models represent you.

The standard is in active adoption, not yet universally mandated. Perplexity and several AI search tools have already confirmed they reference llms.txt. OpenAI and Anthropic are increasingly signalling support for structured web signals — and both browse live web content in their tools. More importantly, llms.txt works in concert with schema markup, well-structured authoritative content, and citation signals that collectively shape how AI models represent your brand. It is not a silver bullet on its own — which is why our service combines it with a full audit and supporting content strategy.

We run a systematic series of prompts across ChatGPT (GPT-4o), Claude, Perplexity, Gemini and Microsoft Copilot. These prompts cover: direct brand queries ("tell me about [company]"), service-based queries ("who provides [service] in [location]"), comparison queries ("compare [company] vs [competitor]"), and trust queries ("is [company] reputable / what do clients say about [company]"). We record every response verbatim, annotate inaccuracies, missing information and positioning gaps, and compile a written audit report with evidence screenshots before we write a single word of your llms.txt.

The audit itself typically takes 5–7 working days. We then present findings in a brief call, gather your input on positioning and priorities, and produce the final llms.txt copy for your approval. From start to a live, deployed file, most clients are completed within 3 weeks. For Managed clients, we schedule a first re-audit at 90 days post-deployment to measure any detectable shifts in AI responses, and then quarterly thereafter.

Deploying an llms.txt file is technically straightforward — it is a plain-text file uploaded to the root directory of your domain, exactly like robots.txt. If you have a web developer or use a CMS like WordPress, Webflow or Shopify, we'll provide the file and step-by-step instructions. If you'd prefer us to handle deployment directly, we can do so with brief FTP or hosting panel access. We will never ask for credentials we do not need and will confirm every action with you before proceeding.

Accuracy today does not guarantee accuracy tomorrow. AI models retrain continuously, and a competitor's PR campaign, a negative review thread, or simply a shift in training data weighting can change how you are represented overnight. Having an llms.txt in place gives you a persistent, authoritative signal that works as a corrective anchor even when other signals change. It also gives you a competitive advantage — if your competitors are not yet using llms.txt and you are, your narrative is the one AI systems have a structured reason to trust. We still recommend the audit even if your current AI representation appears positive, as there are almost always framing gaps or missed opportunities you would not otherwise spot.