Live demo
A pretend bot with the free snippet on it. Try "are you a human?", "ben je een mens?" or switch to timed reminders.
The demo runs entirely in your browser. Nothing you type is sent anywhere.
Is your bot covered? Check before you buy anything
Colorado HB 26-1263 (from 1 Jan 2027)
Covers a public AI system that "primarily simulates human conversation": general-purpose assistants, companion or character bots, open Q&A bots.
Not covered, among others: bots mainly for customer support, shopping, ordering or other commerce; narrow single-topic bots that cannot discuss self-harm or sexual content; internal business tools; developer tools; and chat features inside a larger app that are not built for companionship.
Covered operators must disclose the AI at the first interaction each day, every 3 hours or persistently, and when asked. They also need a suicide/self-harm protocol with crisis referrals, and must report yearly to the Attorney General from 1 July 2027. Duties toward minors go further.
EU AI Act Art. 50(1) (since 2 Aug 2026)
AI systems that interact with people must be designed so that people are told they are dealing with an AI, unless that is obvious from the context.
The Digital Omnibus did not postpone this duty.
This page is general information, not legal advice. Check the text of HB 26-1263 (on leg.colorado.gov) and the Attorney General's rules (final rules not yet published) for your own case.
Free vs. Pro
Free snippet (available now)
- Always-visible "You're chatting with an AI" label
- Fixed, honest answer to "are you human?" questions
- Optional timed mode: first chat of the day, then every 3 hours
- Crisis-keyword hook with hotline links
- English, Dutch, German, French, Spanish
- No dependencies, no tracking; ARIA roles and speaker labels for screen readers
Pro (planned, pre-order)
Pro ships with:
- Server-side log of disclosures and crisis referrals
- Export with the numbers Colorado's annual report asks for
Later, after Pro ships:
- WordPress plugin that wraps your chat widget
- Setup questionnaire that writes your config
- More languages and hotline lists per country
"Shipped" in the refund terms below means the server-side log and the report export are available to buyers. The items under "Later" are not part of that promise.
Pre-orders are not open yet. Sold by ; see the terms and the refund policy.
Payments and privacy: you buy from , which handles your payment and receipts under its own privacy policy. receives the details you enter at checkout, such as your email address, and your order. We use them to deliver the product or refund you, because that is needed for your purchase, and keep order records as long as tax law requires (in the Netherlands, 7 years). Email to see or correct them. See also the privacy page.
Questions
Is a customer-service chatbot covered by Colorado's chatbot law?
Not if it is primarily designed for customer support, customer service, shopping, ordering or other commerce: HB 26-1263 excludes those. It also excludes narrow single-topic bots that cannot discuss self-harm or sexual content, internal business tools, developer tools, and chat features inside a larger app that are not built for companionship.
When does Colorado's AI chatbot disclosure duty start?
On 1 January 2027. Annual reports to the Attorney General start on 1 July 2027.
How often must a covered chatbot in Colorado say it is an AI?
At the start of the user's first interaction each day, at least every 3 hours in a continuous conversation or as a persistent disclosure, and whenever the user asks whether it is human.
Does EU AI Act Article 50 require chatbots to disclose they are AI?
Article 50(1) has applied since 2 August 2026: AI systems that interact with people must be designed so that people are told they are dealing with an AI, unless that is obvious from the context. The Digital Omnibus did not postpone this duty.
Does this snippet make my chatbot compliant?
No. It is a building block: it shows the AI disclosure, answers "are you human?" and links crisis lines. Covered operators also need a full suicide and self-harm protocol, annual reports and, for minors, more. This page is not legal advice.
What does the free snippet do?
It shows an always-visible AI label or timed reminders (first chat of the day, then every 3 hours), gives a fixed answer to "are you human?" in English, Dutch, German, French and Spanish, and shows crisis-line links when a message matches suicide or self-harm keywords. It has no dependencies and no tracking.
Get notified
One email when Pro ships, and one if the Colorado rules change what you need to do.
Sign-up is not open yet.
Privacy
This page is run by . If you sign up, we use your email address only . We do so because you asked us to (your consent), and you can withdraw that at any time.
The address is stored by , the sign-up service, which processes it on our behalf; its own privacy policy says where. We keep it .
Every email we send will have an unsubscribe link. You can also email to see, correct or delete what is stored about you. If you think we handle your data wrongly, you can complain to your data protection authority (in the Netherlands, the Autoriteit Persoonsgegevens). This page itself sets no cookies and has no analytics. See also the privacy page.