22 August 2026: this page said questions that got an answer were never recorded, and for a time that was wrong — a question was recorded as soon as our own guides declined it, before the statute lookup ran. The wording was corrected, and then the behaviour was changed so the original promise is true again. Both steps are logged at corrections.
21 August 2026: the description of what the question box writes its answers from was wrong on this page, and is corrected below. It said answers came from this site’s guides and nothing else; the box has also answered from the Texas statutes since 16 August. The change is logged in full at corrections.
4 August 2026: Google Tag Manager and Google Analytics were added to this site. Before that date the site loaded nothing from any third party and this page said so. It now describes what is actually collected.
The short version
- Google Analytics is installed (property
G-500CJC2HMB), loaded directly, alongside Google Tag Manager (containerGTM-KJG899Q9). Google is the only third party this site loads. - Do Not Track and Global Privacy Control are honoured. If your browser sends either signal, Tag Manager never loads — no script, no request, no cookie, and no tag inside it runs either.
- No geolocation lookup. Nothing here resolves your IP address into a city, region or internet provider.
- No advertising. No ad network is allowed to load, and the site sells nothing.
- No account, no login, nothing to sign up for except an optional email list.
- The question box sends what you type so a language model can answer it — from our guides where we have one, and otherwise from sections of the Texas statutes fetched at the moment you ask. The conversation is never stored and dies with the tab; a question nothing here could answer is recorded — not one the statutes answered, and not one that timed out — and only the question, with no address, no IP and no identifier.
- Nothing is sold, rented or shared.
What Google Tag Manager is, and the honest caveat
Two Google products are involved, and they do different things.
Google Analytics is the measurement tool. It records the ordinary things described below, and it is configured in this site’s own code where the settings are visible and reviewable.
Google Tag Manager is not a measurement tool. It is a loader: a container that can load other tools based on a configuration held in Google’s console rather than in this site’s code.
What loads inside that container can be changed by the site owner at any time without altering a single line of this website. So a sentence here saying “we run exactly these tools” would be a claim this page has no power to keep.
What can be promised, because it is enforced in code and in the site’s security policy rather than in a console, is everything in the next two sections. We would rather tell you the shape of the guarantee accurately than give you a comforting list that could quietly go stale.
The intended use is ordinary: knowing which guides people actually read, so the next ones are written about the right procedures. Nobody here looks at individual readers.
Expect the usual analytics collection: which pages you viewed and in what order, roughly how long you stayed, the approximate location derived from your IP address, your device type, browser and screen size, and the site or search that sent you here. Cookies are set to recognise a returning browser. Those cookies are Google’s, not ours.
The two guarantees that are enforced in code
1. Do Not Track stops everything, permanently
The check sits in front of the container, not inside it. If your browser sends Do Not Track or Global Privacy Control, Tag Manager itself is never requested — so no tag within it can run, including any tag added in future by someone who never read this page.
That placement is deliberate. Configuring privacy on individual tags would leave the guarantee only as durable as the next console change.
2. The security policy caps what can ever load
This site sends a Content-Security-Policy that names the specific origins a browser is permitted to load anything from. Right now that list is Google Tag Manager and Google Analytics, and nothing else. A tag pointed at any other destination — an ad network, a session recorder, a third-party pixel — is refused by your browser, not merely discouraged by us.
Widening that list requires a change to this site’s code, which is a visible, recorded change rather than a console toggle. That is the part worth trusting: the set of companies that can receive anything about your visit cannot be expanded from a web console.
To be straight about the limit of this: Tag Manager is permitted to run custom code of its own inside the page. So the guarantee is about who your browser may talk to, not about what the code from those two Google origins may do once it is running.
Turning it off
- Enable Do Not Track or Global Privacy Control in your browser. This site checks for both before loading anything. Most browsers have it in privacy settings; some, including Brave and DuckDuckGo’s browser, send it by default.
- Use a content blocker. uBlock Origin and similar block Tag Manager outright. Nothing on this site breaks if you do — every page works with all scripts blocked.
- Install Google’s own Analytics opt-out browser add-on, which disables Google Analytics across every site.
You can verify this. Open your browser’s developer tools, look at the network tab, and reload a page. With a Do Not Track signal set there should be no request to any Google domain at all.
What else is collected, and by whom
Server logs
This site is hosted on Netlify. Like every web host, Netlify processes the technical information required to serve a page — IP address, the page requested, the time, and your browser’s user-agent string. We do not add to it and we do not use it to build any picture of individual readers. Netlify’s handling of that data is governed by its own privacy policy.
If you would rather not appear in any web server’s logs at all, that is a general property of the web rather than of this site, and a VPN or Tor is the usual answer.
The email list, if you join it
If you enter an email address on the homepage, that address is stored by Netlify Forms. It is used to send occasional notices about new guides and corrections, and for nothing else.
- We do not sell, rent, share or trade the list.
- We do not use it to profile you, and we know nothing about you beyond the address you typed.
- Ask and it will be deleted. No explanation needed.
The form does not require your name, and we would rather you did not give us one. An email address is enough.
The chatbot, and what it sends
What you type into the question box is sent to this site, and on to Anthropic’s API, so that a language model can write you an answer. The last few messages of the conversation go with it, so that a follow-up question makes sense. This changed in August 2026: the box used to answer from a table inside the page and send only the questions it could not place. It now sends nearly everything, and this page would rather say so plainly than describe a version of the box that no longer exists.
If your question is a civil-law one this site has no guide for, it also goes on to find the Texas statute that governs it. The Legislature’s server is contacted by this site, not by your browser — it never sees your address, and it is asked for a chapter of the law, not for anything about you.
Two things are still answered without anything being transmitted: a bare rule number like 21c or 251, which is looked up in the page itself, and anything at all if the call fails — if the service is down or unreachable, the box falls back to that same built-in table rather than failing shut.
The answer you read is written by the model, and which text it was allowed to write from depends on which of the two paths above answered you. When the answer comes from our guides, the model may use that guide’s text and no other material: it is instructed not to state any rule, deadline, cost or form name that is not in front of it, and to say plainly that this site does not cover something rather than reach for an answer. Every link and every rule citation shown to you there is hand-written and checked against the official source — the model picks which of our pages and which of our rules apply, it does not compose them. When the answer comes from the statutes instead, it is written from sections fetched from the Legislature at the moment you asked, which nobody here has read; it quotes their conditions rather than paraphrasing them, and it is labelled as a statute answer every time so you can tell the two apart. On both paths the model is instructed never to tell you what any of it means for your own case.
We do not keep your conversation. It lives in the page while the box is open and is gone when you close the tab — not in a cookie, not in local storage, not on a server, and not connected to anything else you did here. We keep no transcript and could not produce one.
The single exception: a question this site’s own guides could not answer is recorded, so a person can read it and write the missing guide. What is kept is the question and nothing else — no email address, no name, no account, no IP address, no identifier of any kind, and not the rest of the conversation. It is truncated, and it is not connected to a session, so two questions from the same person cannot be told apart from two questions by strangers.
Questions that got an answer are not recorded — not by the guides and not from the statutes. Earlier on 22 August 2026 that was not true: the recording happened the moment the guides declined, before the statute lookup ran, so questions the statutes answered were being written down too. That is fixed. The record is now made at the one point where you are actually told this site cannot help.
A question that simply took too long is not recorded either. Running out of time is not the same as having no answer, and it would be dishonest to file it as one.
Your browser sends it, not our server. Two things follow, and both are to your advantage. If you close the tab before the answer finishes, nothing is recorded at all. And because it happens in the page rather than behind it, you can watch it in your browser’s network tab like anything else here.
Anthropic, who run the model, receive your question in order to answer it and handle it under their own API terms. We have no arrangement with them about your words beyond that, and we do not send them anything identifying you.
So: do not type your cause number, your county, your court, or anybody’s name into it — and never a child’s. The box does not need any of that to answer, and every one of those details is something that need never have left your hands.
We record the failures because the alternative is worse: a box that keeps failing the same way, and nobody ever finds out. Records are deleted once the gap they describe has been written up.
The corrections form, if you use it
The form on Corrections is stored by Netlify Forms, the same as the email list. It asks which page is wrong and what is wrong with it. The email field is optional and an anonymous report is as useful to us as a signed one.
It deliberately does not ask for your cause number, your county, your court or the names of anyone in your case, and the form says so above the fields. If you send them anyway we will not use them, but we would rather never hold them: readers of this site are in contested proceedings, and the safest record of somebody’s case is the one that was never created.
Reports are kept while the page they concern is still live, so a correction can be traced back to what prompted it. Ask and yours will be deleted.
Links to other sites
Following a link away from here — to TexasLawHelp, a county clerk, a legal aid provider, the Legislature’s statute site — takes you somewhere with its own practices. We have no control over and no visibility into what those sites do.
What is deliberately absent
An earlier placeholder at this address called a geolocation service that resolved each visitor’s IP address into a city, region and internet provider, writing the result into a hidden form. The sibling site carried the same code and it was removed on 3 August 2026.
None of it was malicious — that combination is a common template and most people who ship it never look at what it does. But a site whose readers are in contested litigation has no business building a log of where they read it from. It is not here, it is blocked at the Content-Security-Policy level, and it is not coming back.
Also absent: web fonts, content delivery networks, embedded video, social widgets, session recording, heatmaps, and A/B testing tools — and each of those is blocked by the security policy described above, not merely unused. Apart from Google Tag Manager and Google Analytics, every file a page loads comes from this domain.
If you are in an unsafe situation
This site cannot see who you are, but the device you are reading on can, and so can anyone with access to it. If someone else has access to your phone or computer, consider a private browsing window, and clear your history afterwards.
If a shared account, shared phone plan or shared computer is part of what you are dealing with, the National Domestic Violence Hotline publishes guidance on covering your tracks online, and it is better than anything we could write here.
Changes
If what this page describes changes again, the page changes with it and carries a dated note at the top saying what moved, as it does now. A privacy policy that is quietly rewritten is worth nothing.
Last reviewed 22 August 2026.