Abuse policy
Warnings, 24-hour suspensions and what counts.
- A web page my agent read contained insults. Will I get a strike?No. Only the newest message a person wrote is checked, never tool output such as fetched pages or file contents.FAQ
- Can I ask the assistant to help build a profanity filter?Yes. Asking about a word, translating it or building a filter for it doesn't count as abuse.FAQ
- Do I need to do anything when my 24-hour suspension ends?No. Your account works again on its own when the 24 hours end. Don't retry API calls before resets_at.FAQ
- Does the abuse policy apply in every language?The published policy doesn't limit itself to one language: insulting or cursing at the assistant isn't allowed.FAQ
- How do I appeal an abuse-policy strike or suspension?Contact Support with your Account ID and the time of the request. If the appeal is accepted, the suspension is lifted early.How-to
- How do I protect my app from abuse-policy suspensions?Moderate end-user input before sending it, log the warning header, and block users who trigger it.How-to
- How long does a strike stay on my account?Strikes are counted within a 30-day window. The suspension itself lasts 24 hours and lifts automatically.FAQ
- I pasted logs or code that contain swear words. Is that a strike?No. Quoted text such as code, logs, pasted errors, string literals and quotes doesn't count.FAQ
- If one team member gets suspended, does the whole team stop?A suspension applies to the account that wrote the messages, or to the key owner's account for API text. Use separate keys and moderation.FAQ
- My app sends users' text to the API. Who gets the strike?You do. Text your app sends through your API key counts against your account, exactly as if you wrote it.Policy
- What counts as abuse under the abuse policy?Only insults or cursing aimed at the assistant. Swearing at a bug, quoted text, or asking about a word don't count.Reference
- What do the abuse warning headers and suspension error look like?Warnings add x-clatter-notice and x-clatter-notice-text headers. Suspensions return 403 account_suspended with resets_at and Retry-After.Reference
- What happens when I get a warning or a strike?A first strike in 30 days is a warning; your request still works. A second within 30 days pauses the account for 24 hours.FAQ
- What is the Clatter Labs abuse policy?Insulting or cursing at the assistant earns a warning; a second time within 30 days pauses your account for 24 hours on every product.Policy
- What still works during an abuse-policy suspension?Support stays reachable. Credits, plan, keys and data are untouched; all products resume after 24 hours.FAQ
- Which messages does the abuse check look at?Only the newest message a person wrote. Never tool output, system prompts or the model's own replies.FAQ
- Which products does the abuse policy cover?Line CLI, the Model API, the Agent API, the Console Playground and Cosmos Agent, including on Telegram.Reference
- Why did the assistant warn me about rude words?Being rude or swearing at the assistant isn't allowed. The first time is a warning; a second time within 30 days pauses your account for a day.FAQ
- Will I get a strike for swearing at my code or a bug?No. Frustration aimed at a bug, a build, a tool or a third party doesn't count. Only insults aimed at the assistant do.FAQ
Still need help?
Tried the steps above and still stuck? Email us. We usually reply within the first hour, and always within 24 hours. A full support team is with you.
Infrastructure questions, or permanent deletion of your data
infrastructure@clatterlabs.xyzFor deleting your whole account or the data of one service.