EsqHarness | Legal AI risk control for U.S. lawyers
Your AI Draft Is Only as Reliable as the Cases It Cites.
Find out whether your firm’s use of ChatGPT, Claude and other AI tools exposes you to citation errors, court-rule violations and avoidable professional risk.
- Identify gaps in your firm’s AI safeguards.
- Receive a personalized AI Practice Risk Assessment.
- Discover practical steps to strengthen AI oversight.
About 4 minutes · Free assessment · No commitment
See How EsqHarness WorksWhat you can rely on
Built for Lawyers Who Use AI, and Remain Responsible for the Work
Citation and quotation checks
Quotations and case citations are checked against CourtListener, the case law database of the nonprofit Free Law Project. A made-up case comes back as not found.
Inside Claude and ChatGPT
EsqHarness is added as one link in the settings of Claude or ChatGPT plans that allow custom connectors. Nothing to install.
Court AI rules, with sources
California rules, ethics opinions and federal standing orders on AI, each linked to its source and checked by a person. The source pages are checked for changes every week.
Confidentiality
Text you check is sent to our server only to run the check and is not stored. We keep the counts, the citations found and the result. You can remove client names first.
You still review the work
A check finds what the database holds. A “found” result does not mean the case supports your point. You read what you cite and remain responsible for every filing.
The problem
AI Can Draft in Seconds. You’re Still Responsible for What Gets Filed.
1
Fabricated or inaccurate citations
AI-generated legal research may contain citations and quotations that need independent verification.
EsqHarness: checks each case and quotation before the draft reaches you. Working today
2
Court-specific requirements
AI disclosure and certification requirements differ across courts and judges.
EsqHarness: looks up the AI rules of your courts and drafts the disclosure a court asks for. California today Other states planned
3
Inconsistent firm safeguards
Without a written process, each attorney may use AI in a different way.
EsqHarness: one firm AI use policy, approved by one of your attorneys, applied in every chat, with a monthly one-page record. Working today
21 of 23
quotations in an opening brief were fabricated by AI. In Noland v. Land of the Free, L.P. (2025) 114 Cal.App.5th 426, the California Court of Appeal sanctioned the attorney $10,000 and referred the matter to the State Bar. Read the published opinion.
See it work
Your AI Checks Its Own Work Before You Do
Tell us your courts
We confirm the courts your firm files in, the attorneys covered and that your Claude or ChatGPT plan allows custom connectors.
Approve your harness
We send your firm’s AI use policy and AI instructions. One attorney reads them and clicks Approve.
Paste one link
In Claude or ChatGPT, open the settings, add your EsqHarness link and switch it on. Setup takes about 3 minutes.
How the pieces fit
One Platform, One Product, One Free Starting Point
ESQ CoPilot
The platform for U.S. lawyers: courses, the lawyer directory and practice tools, including EsqHarness.
EsqHarness
The legal AI risk-control product. It runs inside Claude and ChatGPT and checks citations, court rules and your firm policy as you work.
AI Practice Risk Score
A free assessment of your firm’s AI safeguards. It is not EsqHarness and does not check any document.
Availability today
- For solo and small firms (1 to 10 attorneys) using ChatGPT or Claude, filing in California state or federal courts.
- New firms are added by request while new terms are prepared. No price or trial is offered on this page right now.
- Founding Firm trials already started continue as agreed.
Built by
Christian Kameir
Author of the eBook Ethics-Safe AI Use for Law Firms (Decentralized Publishing LLC, 2025), developer and operator of a production legal AI platform used in legal practice. He is a law graduate of the University of Muenster, Germany, and is not admitted to practice law in any U.S. jurisdiction. More on the faculty page
Questions?
Email support@alcbar.com. ESQ CoPilot is not a law firm. It provides compliance tools and education, does not give legal advice and creates no attorney-client relationship.
Questions
Common Questions
Is the free Risk Score the same as EsqHarness?
No. The Risk Score is a free set of questions about your firm’s AI safeguards. EsqHarness is the product that checks citations and court rules inside your AI tools.
What is a harness?
The instructions, safeguards and checks that control how an AI tool works. EsqHarness turns current state rules, ethics duties, court orders and your firm’s policy into a harness for your AI and keeps it up to date.
Do I need new software?
No. You add one link in the settings of Claude or ChatGPT. Setup takes about 3 minutes.
Does it guarantee I will not be sanctioned?
No. It catches made-up quotations and citations that the case law database can test, and reminds you of your court’s rules. You still read what you cite and remain responsible for every filing.
Which courts do you cover?
California state courts and the federal courts in California first. Other states follow.
Is this legal advice?
No. ESQ CoPilot is not a law firm. It provides compliance tools and education, does not act as your lawyer, and creates no attorney-client relationship.
What happens to my Risk Score answers?
Your score appears on screen without sign-up. If you ask for the full report by email, we keep your name, email, firm details and scores. See the Privacy Policy.
Who is Decentralized Publishing LLC?
The company that operates ESQ CoPilot. Its name appears at checkout and on your card statement. Address: 92 Corporate Park, Ste C #1083, Irvine, CA 92606.
Start With Where Your Firm Stands Today
Twelve questions. Your score appears on screen right away.
EsqHarness is a service of ESQ CoPilot. ESQ CoPilot is not a law firm, does not give legal advice and creates no attorney-client relationship.