These terms cover your use of habits.chat. Right now that means this website and the early access waitlist; once accounts open, they cover the application, the chat features and the MCP server too.

The waitlist

Joining the waitlist creates no obligation on either side. It doesn't guarantee access, doesn't reserve a place in a queue we've promised to honour in order, and costs nothing. We may contact you about early access and product progress; you can leave at any time using the unsubscribe link in any email.

Your account

You're responsible for keeping your credentials secure and for activity under your account, including calls made by any AI assistant you've connected. You must be at least 16 to use the service. One person per account.

Your data

Your habits and check-ins belong to you. You grant us only the licence needed to operate the service — storing it, displaying it back to you, processing it to answer your questions. We don't claim ownership, we don't use it to train models, and we don't sell it. You can export it or delete it at any time.

The AI features, and what they are not

habits.chat is a tracking tool. It is not a medical device and does not provide medical, psychological, addiction or health advice. Several of the things people track here — quitting smoking or vaping, reducing alcohol, sleep, food, and everything in the autism section — sit close to real medical questions. For those, talk to a doctor or a qualified professional. Nothing the app or the chat says is a substitute for that, and it should not be used to make decisions about medication, treatment or withdrawal.

Specifically, and without limiting the above: it does not diagnose or screen for autism, ADHD, a tic disorder, a sleep disorder, a food allergy, coeliac disease, IBS or anything else; it does not assess or interpret anyone's behavior; and it does not recommend or evaluate any therapy, intervention, diet or medication. Counts and averages it reports are arithmetic over entries you made, not clinical findings, and are not evidence of anything on their own.

Food, allergies and elimination diets

This one carries a real risk, so it gets its own paragraph. Do not use the service to test whether you or anyone else reacts to a food you believe you may be allergic to. Deliberately eating a suspected or known allergen is an oral food challenge, and it belongs in a clinical setting with adrenaline and trained staff present, because a reaction can become life-threatening within minutes. Any reaction involving breathing, swelling of the lips, tongue or throat, or faintness is a medical emergency — adrenaline if it has been prescribed, then emergency services. Nothing in the service is an allergy action plan or a substitute for one.

The food pages support a protocol you are already running, ideally with a GP or dietitian. They do not design a diet, do not decide what to eliminate, and do not turn a pattern in your log into a diagnosis. Elimination and restriction carry their own risks — nutritional deficiency, and in children the possibility of creating an allergy that wasn't there — which is why the pages point at a dietitian rather than around one. You remain responsible for decisions about what you or anyone in your care eats.

Recording information about another person

If you use the service to record what's happening for someone you care for — including your own child — you're responsible for having the standing to do so, and the account and its contents remain yours rather than theirs. Record only what you're comfortable holding. The service is a personal notebook and is not built or certified as a clinical, educational or safeguarding record system, so please don't rely on it as one.

The chat is powered by a third-party language model. It selects which operations to run against your record; the figures it reports are computed by the application rather than generated by the model. It can still misunderstand a request or call the wrong tool, which is why every reply shows what it did and why writes are confirmed and reversible. Check anything that matters.

Connected assistants

You may connect MCP clients to your account using tokens you create. You're responsible for which clients you connect and what scopes you grant them, and for revoking access when you no longer want it. Data that reaches a third-party client is then governed by that vendor's terms. We log every call so you can see what was done.

Acceptable use

Don't use the service to break the law or to harm anyone.
Don't attempt to access other people's accounts or data.
Don't scrape, resell or redistribute the service, and don't hammer the API in a way that degrades it for other people.
Don't upload other people's personal data as habit content.

Billing

Paid plans are billed in advance, monthly or yearly. Founding accounts keep the rate they signed up at for as long as the account remains open. You can cancel at any time; cancellation takes effect at the end of the current period and drops the account to the free tier rather than deleting anything. Yearly plans cancelled mid-term are refunded pro rata.

Availability

We'll try to keep it up and will give notice of planned downtime where we can, but the service is provided as-is without a formal uptime guarantee during early access. If we ever discontinue it, we'll give at least 60 days' notice and keep exports working throughout.

Liability

To the extent permitted by law, our liability is limited to the amount you paid us in the twelve months before the claim. We aren't liable for indirect or consequential losses, or for decisions made on the basis of what the app or the chat told you.

Changes

We may update these terms. Material changes will be emailed to account holders before they take effect. Continuing to use the service after that means you accept the new version.

Contact

Questions about these terms: hello@habits.chat.

Terms of Service