llms.txt for law firms
Does a law firm need an llms.txt file?
What llms.txt is, who says they read it and who says they don't, why we keep one on our own site, and what does more for a law firm.
By Santiago Alvarez, Founder, Ad Hoc Digital
Last updated
The short answer
If you read one part of this page, read this.
No. llms.txt is a proposal for a plain text summary of a website that AI agents can read, and Google says Google Search ignores it: having one will "neither harm nor help" a site's visibility, AI Overviews and AI Mode included. None of the AI companies whose crawler pages we read says its search uses one either.
It isn't harmful, and our own site has one. We keep it as a tidy index of our pages for any agent that looks, not because we expect it to get us named in ChatGPT. The risk for a law firm is a second copy of its facts that goes stale.
If someone is selling you an llms.txt as AI search optimization, the money does more on practice pages, listings and reviews. We work with law firms across the US and Canada, and also in Australia and the UK.
What it is
llms.txt is a proposed Markdown file that summarizes a site for AI agents.
Jeremy Howard published the proposal on September 3, 2024 and updated it to a second version in August 2026. It's a convention some sites follow, not a standard any search engine has adopted.
- The format
A Markdown file at /llms.txt. Only a heading with the site's name is required. After that come a short quoted summary, optional notes, and sections of links, each link with a line saying what's on the page. A section called Optional holds links an agent can skip.
- What it's for
The proposal says it's used on demand, when an agent needs information while helping someone, and that it expected the file to matter for that rather than for training. Its heaviest use is in software documentation, where coding tools follow it to find reference pages.
- What version 2 added
Clean Markdown copies of pages at the same address with .md added, and link tags that point an agent from a page to its Markdown copy and to the llms.txt that covers it.
- What it isn't
It isn't robots.txt: it doesn't allow or block anything. It isn't a sitemap either. The proposal itself describes it as a curated overview, not a list of every page.
Who reads it
Google Search says it ignores llms.txt, and no AI search crawler says it reads one.
The clearest statement comes from Google. The AI companies are quieter: they publish llms.txt files for their own developer docs without saying their search tools read yours.
| Who | What they say | What it means for a law firm |
|---|---|---|
| Google Search (incl. AI Overviews and AI Mode) | Google Search ignores llms.txt; it neither harms nor helps visibility | No effect on Google, either way |
| Google Chrome Lighthouse | An experimental agentic browsing check looks for llms.txt at the domain root | A browser audit for agents, not a ranking factor |
| OpenAI | Its crawler page links OpenAI's own docs llms.txt; it says nothing about reading yours | OAI-SearchBot access is what matters for ChatGPT search |
| Perplexity | Same: its crawler docs link its own llms.txt, nothing about reading sites' files | PerplexityBot access is what matters |
| Microsoft Bing and Copilot | Bing's Webmaster Guidelines don't mention llms.txt | Bing Webmaster Tools, sitemaps and IndexNow matter |
Google added its note on June 15, 2026, saying it wanted to answer questions from the community. It also says Google may crawl and index all kinds of files on a site, and that being crawled doesn't mean a file gets special treatment. So if you see Google fetching your llms.txt in server logs, that's not a signal.
The Lighthouse check is real, but Chrome labels the whole agentic browsing category experimental, needing Chrome 150 or later, with a fraction of checks passed instead of a score out of 100. It's about agents using a site, not about which lawyer an AI recommends.
Harmless or not
The file can't hurt your rankings, but it can repeat old facts.
Google says it neither helps nor harms. The risk for a law firm is the same as any forgotten page: it says something that stopped being true.
An llms.txt usually restates the firm's name, practice areas, lawyers, cities and phone number. Then a partner leaves or the office moves, the website gets updated, and nobody remembers the text file. Any tool that reads it gets the old version.
We found exactly that on our own site in October. Our llms.txt linked to a page that no longer existed and carried a claim with no source behind it, so we rewrote it to match the site. On a client site we took over, the existing llms.txt still spelled the office address the old way and listed languages the firm hadn't confirmed; we flagged it for a rewrite from the firm's master record.
It's also public. Anyone can open it, so in our reading it's marketing like any other page: the same bar rules on claims, results and specialist wording apply to what it says.
Our own file
We keep an llms.txt on our site as an index of our pages, and we don't count on it for anything.
It's at /llms.txt on our site, and saying so honestly is part of this answer. Here's what's in it and why.
- What's in it
Our name, a one-paragraph description that includes the line about where our clients are, then links with one-line descriptions: the main pages, each service, each practice page, every guide in this library grouped by topic, and the two founders.
- How it stays current
The guides section is rebuilt from the same data that runs the guides index, every time a guide is added, and a test fails if any guide is missing from it. That's the only reason we trust it not to drift.
- Why we keep it
It costs almost nothing once it updates itself, and an agent asked about law firm marketing gets a clean map of our pages if it looks. We don't measure it, report on it or expect it to change how often we're named.
- What we do for clients instead
We don't add one to client sites as an AI search step. Our site audits note whether a firm has one and mark it "not a lever". Where a firm already has one, we make sure it matches the website.
What does more
Five things move a firm's odds in AI answers more than any text file.
Each of these changes what Google, Bing or the AI tools can actually read about the firm. Do them before thinking about llms.txt.
Let the search crawlers in
Googlebot, Bingbot, OAI-SearchBot and PerplexityBot, checked in robots.txt and at the CDN or firewall. A blocked crawler can't read any file, llms.txt included.
Check Google's AI setting
Google requires sites to be included under Search generative AI in Search Console to appear in its AI features. Include is the default; confirm nobody changed it.
Write a page per case type
The answer in the opening lines, the local court and process, the law cited to official sources. Our guide on AI search vs SEO explains why Google's fan-out searches reward this.
Fix the listings AI tools cite
Business Profile, Bing Places, Yelp and the legal directories, all matching the website. Our guide on where AI tools find lawyers shows how to find yours.
Keep reviews and labels accurate
Steady reviews from every client, and ordinary structured data that matches the visible page. Our structured data guide covers which types fit a law firm.
If you want one anyway
If you publish one, write it from the same facts as your website and update them together.
There's nothing wrong with having one. These rules keep it from becoming the stale copy described above.
| Part | What to put | Rule |
|---|---|---|
| Heading | The firm's exact name | Identical to the website, Business Profile and bar listing |
| Quoted summary | Practice areas, cities served, and how to reach the firm | Only facts the website shows; the real phone number |
| Practice section | One link per practice or case type page, with a one-line description | Only pages that exist and that you want to be known for |
| Lawyers section | One link per lawyer bio | Remove a lawyer the day they leave |
| Optional section | Articles and guides | Skip anything out of date |
Why it gets sold
llms.txt gets sold because it's easy to deliver, not because it moves AI answers.
A file you can see at a web address looks like work done. That makes it an easy line item in an AI search package.
Google's guidance on third-party SEO advice, which names AEO and GEO tools directly, says good advice either cites official guidance or labels itself as opinion, and that Google doesn't evaluate or approve outside services. Apply that here: ask the vendor which AI tool reads the file and which of that company's pages says so. For Google, its own page says it doesn't.
More on how estate planning clients look for a lawyer in estate planning marketing.
What we see
In our client work, the firms AI tools name got there without an llms.txt.
None of our AI search results depended on the file. They came from the slower work it's often sold in place of.
Our clearest result is a family law firm that now shows up first in ChatGPT: 50 to 100 extra calls a month over the last three months. Another moved from family law into estate planning, and prospective clients started finding it inside ChatGPT and similar tools. Both came from pages, listings and reviews, the core of our AI search work.
When we audit a firm's site, an llms.txt question takes one line of the report. The rest goes to crawler access, Search Console, the pages and the listings, and that's where firms' budgets should go too. The websites we build carry the parts Google does read: clear pages, accurate structured data and nothing blocking the crawlers.
Common mistakes
Where firms go wrong.
We see these when a firm has already bought into llms.txt as an AI fix.
Treating it as an AI search step
Google says Google Search ignores it, and no AI search crawler documents reading it. It belongs at the bottom of the list, if it's on the list at all.
Letting it drift from the website
A departed lawyer or old phone number in llms.txt is one more wrong source. Update it whenever the website changes, or don't publish one.
Confusing it with robots.txt
llms.txt blocks nothing. If the goal is to control AI crawlers, that's robots.txt and the CDN, crawler by crawler.
Paying for Markdown copies of every page
The proposal suggests them for agents, and Google says it doesn't need them. For a law firm site, the HTML pages are what search engines and AI features read.
Real results
What this looked like for real firms.
Two firms clients now find inside ChatGPT, neither of which needed an llms.txt to get there.
Identifying details are anonymized to protect our clients. Individual result, not a promise or prediction of any specific outcome for your firm.
FAQ
Questions lawyers ask us.
Straight answers to the questions that come up most.
Does Google use llms.txt for AI Overviews or AI Mode?
No. Google's guide to its generative AI features says Google Search ignores llms.txt, and that creating one will neither harm nor help a site's visibility or rankings. Its AI features page says you don't need AI text files or special markup. Google says it's fine to keep one for other services that use it.
Does ChatGPT read llms.txt?
OpenAI doesn't say so. Its crawler page describes OAI-SearchBot, GPTBot and ChatGPT-User and how robots.txt controls them, and it links an llms.txt for OpenAI's own docs, but nothing says ChatGPT search reads a site's llms.txt. What OpenAI does say matters is allowing OAI-SearchBot. Our ChatGPT guide covers that.
Can an llms.txt file hurt our website?
Not in Google, which says it neither harms nor helps. The practical risk is accuracy: if it lists an old address, a lawyer who left or a practice you dropped, any tool that reads it repeats that. Keep it in step with the website, and apply the same bar rules to its claims.
Why does your own site have one if it doesn't help?
Because it costs almost nothing to keep accurate, and it gives any agent that looks a clean map of our pages. Its guide list rebuilds from the same data as our guides index, so it can't drift. We don't expect it to change how often AI tools name us, and we don't sell it to clients as a step.
Should we add llms.txt to our law firm website?
Only after the work that matters is done, and only if someone will update it with the website. It's optional and Google ignores it. If you have a limited budget, crawler access, practice pages, listings and reviews all come first.
What's the difference between llms.txt and robots.txt?
robots.txt tells crawlers what they may and may not fetch, and the AI companies document how their crawlers follow it. llms.txt is a proposed summary of the site for agents; it allows or blocks nothing. If you want to control AI crawlers, robots.txt and the CDN settings are where that happens.
Can you check whether our site's llms.txt is accurate?
Yes. It's one line in the site audits we run, next to crawler access, Search Console and listings, which matter far more. If you'd like us to look at your site and what AI tools say about your firm, schedule a consultation.
Where we do this
The services this guide touches.
What this looks like when we run it for a firm, with a demo for your practice on each page.
Sources
Where these facts come from.
Official pages we read when writing this page. Platforms and rules change, so check the current version before you act on any of it. This is marketing guidance, not legal advice.
- llmstxt.org, The /llms.txt file, v2 (proposal)
- Google Search Central, Optimizing your website for generative AI features on Google Search
- Google Search Central, Latest documentation updates (June 15, 2026 llms.txt note)
- Google Search Central, AI features and your website
- Google Search Central, Guidance on third-party SEO tools, services and advice
- Chrome for Developers, Lighthouse agentic browsing scoring
- OpenAI, Overview of OpenAI crawlers
- Perplexity Docs, Perplexity Crawlers
- Bing Webmaster Tools, Webmaster Guidelines
Written by
Founder of Ad Hoc Digital. Leads strategy and works directly with every client firm on AI search, Local Services Ads, Google Ads and Meta ads.
More about SantiagoWant a second pair of eyes on this?
Book a free 30-minute call. Tell us how cases come in today, and we'll tell you straight what we'd change, and whether we can help.