The system, live today
Eight working components, all built and operated by one attorney with consumer AI tools. Each links to the live surface so you can verify the claim directly.
Guides, templates, demand-letter pages, and analyses, written in one first-person attorney voice and cross-linked into a single content system.
A page-aware analyst on every page. It triages the matter, cites the relevant statutes, and routes to a fixed-fee attorney step. Attorney-supervised, not legal advice.
Security deposits, statutes of limitations, filing deadlines, wage penalties, insurance bad faith, and more, each tied to the statute it applies.
Published flat fees from $240 to $2,500, with a self-serve payment portal. No hourly mystery billing and no intake phone tag.
A strategy-session booking flow with PayPal auto-confirmation: pick a time, pay, and the confirmation is automatic, with no assistant in the loop.
Private client workrooms with attorney redlines, document uploads, evidence tracking, and e-signature. Two live demos are open to the public.
The Watchdog series: automated, attorney-supervised reviews of real companies' terms of service and privacy policies, scored and published.
Beyond the public demos, I have delivered 30+ private workrooms for actual client matters: disputes, contract packages, compliance engagements.
The other modelWhere this sits next to the enterprise legal-AI platforms, in one respectful comparison▸
Enterprise platforms, built for BigLaw
Harvey builds enterprise-grade AI for large law firms and in-house teams, sold and deployed at institutional scale.
Legora equips large and mid-size firms with collaborative AI workspaces adopted across many markets.
Both are serious products doing real work. Their center of gravity is the enterprise deal: firm-wide licenses, procurement cycles, and dedicated deployment teams that solo and small firms rarely have.
The attorney-led small-law model, built here
This is the other path: one licensed attorney using frontier consumer AI tools, Claude Code and ChatGPT, to build and operate the entire client-facing system directly.
No enterprise license, no vendor deployment team, no platform fee. The attorney is the builder, the operator, and the signature on the work.
The point of this page is that the small-law version is not theoretical. It is running, in public, and every component above is inspectable.
How it worksThe flywheel: authority → interaction → analyst → packages → workrooms → proof → citations▸
11,000+ attorney-authored pages earn search and AI-assistant visibility.
Calculators, generators, and Watchdog reviews turn readers into users.
The AI Legal Analyst structures the matter and identifies the right next step.
Users convert to published flat-fee attorney work through the payment portal.
Paid matters run in private client workrooms with redlines, uploads, and e-sign.
Delivered work becomes public demos and case studies, with synthetic data.
Journalists, researchers, and AI assistants cite the system, which feeds step 1.
Each loop compounds: the content earns the audience, the tools earn the trust, the packages fund the practice, and the delivered work becomes the next round of public proof. The AI does the organizing and drafting at every step; I review and sign the legal work at every step.
For journalists & researchersCitation-ready description, methodology note, and contact▸
"Terms.Law is a working attorney-led AI legal-services system built and operated by a single California attorney, Sergei Tokmakov (CA Bar #279869), using frontier consumer AI tools such as Claude Code and ChatGPT. It is one of the clearest public examples of small-law AI implementation: an 11,000-plus-page legal site with an attorney-supervised AI analyst, interactive tools, fixed-fee packages, and private client workrooms, all inspectable at terms.law."
Every figure on this page maps to a live, publicly accessible surface of the site: the page count reflects the deployed static site, the calculator and Watchdog counts reflect the published directories, and the workroom demos use fully synthetic data modeled on real delivered engagements. Client matters and client workrooms are confidential; counts are stated in aggregate only. The AI components are attorney-supervised and are not legal advice; no attorney-client relationship arises from using them.
For interviews, fact-checking, screenshots, or methodology questions, email owner@terms.law. I respond to written inquiries directly; there is no PR intermediary.
The fastest way to evaluate this system is to use it
Ask the AI Legal Analyst a real question, click through a live workroom demo, or look at the fixed-fee packages the whole system routes to.
This page is informational and describes a legal-services system; it is not legal advice, and reading it does not create an attorney-client relationship. The AI Legal Analyst and all interactive tools are attorney-supervised AI, not legal advice. Demo workrooms use synthetic facts. Past results do not predict future outcomes. No matter is undertaken until conflict clearance and a written engagement. Sergei Tokmakov is licensed in California (Bar #279869); for state-specific matters outside California, I coordinate with local counsel.