Terms of service

Effective 22 September 2026. Last updated 23 September 2026; the one-Free-organization rule in section 2 applies from 8 October 2026.

These terms govern the use of Chat Guard, the chat moderation service at chatguard.dev, operated by Kirill Tolmachev pr Beograd (Belgrade, Serbia; "Chat Guard", "we"). By creating an organization, adding an API key to a game or server, or sending a request to the API, you ("the customer", "you") agree to them on behalf of the studio or company you represent. If you do not agree, do not use the service.

1. The service

Chat Guard receives chat messages from your game servers or client builds, returns moderation verdicts (categories, severity, target and a recommended action), and offers a dashboard to manage projects, API keys, thresholds, rules and logs. The Unity package and the HTTP API are the documented ways to use it. Verdicts are calibrated probabilities produced by a machine-learning model; they can be wrong. You decide what your game does with them, and you remain responsible for the moderation policy you apply to your players.

2. Accounts and organizations

You sign in with a Google or GitHub account. An organization holds projects, members and billing. Owners and admins may invite members; you are responsible for everyone you invite and for keeping your sign-in accounts secure. Tell us at support@chatguard.dev if you believe an account or key has been compromised.

Each person or company may own one organization on the Free plan. Creating extra accounts or organizations to get more free usage than one Free plan gives, or moving keys between organizations for that purpose, breaks these terms; we may then suspend those organizations or ask you to move to a paid plan.

3. API keys

Server keys (cg_live_) must stay on servers you control. Publishable keys (cg_pub_) may ship in client builds and carry the limits described in the documentation. Test keys are for development. Requests made with your keys are yours, whoever sent them, until you revoke the key in the dashboard. Do not share keys between organizations or resell access to the API.

4. Acceptable use

You may not use Chat Guard to

• process content you have no right to process, or in breach of the privacy notice you give your players;
• probe, overload or interfere with the service, or circumvent quotas, rate limits or key restrictions;
• build a competing moderation service by systematically collecting verdicts;
• moderate chat for services that are themselves unlawful.

We may suspend keys or organizations that break these rules, and will tell you when we do unless the law prevents it.

5. Plans, quotas and fees

Each organization is on a plan (Free, Indie, Studio or Enterprise) with an included number of model verdicts per rolling 30 days and, on paid plans, an overage rate. Current prices and quotas are on the pricing section of chatguard.dev and in the dashboard; they apply as shown there at the time of use. What counts as a metered model verdict is defined in the API reference.

Paid plans are sold through Polar Software Inc. (polar.sh), our merchant of record. Polar charges your payment method, issues invoices and handles taxes and refunds under its own terms, which you accept at checkout. Subscriptions renew monthly until cancelled; cancellation takes effect at the end of the current period, and overage accrued during a period is billed with it. We may change prices with at least 30 days' notice by email or in the dashboard; changes apply from your next renewal.

Above the included quota, Free organizations receive local-filter answers instead of model verdicts; paid organizations are billed the overage rate.

6. Availability and support

We run the service on a best-effort basis and aim for high availability, but we do not promise a particular uptime unless an Enterprise agreement says so. When the model is unreachable the API answers with a local dictionary filter and marks the response as degraded, as described in the documentation. We may change or retire features with reasonable notice, and may interrupt the service for maintenance or security reasons.

Support is by email at support@chatguard.dev. Response times depend on the plan; Enterprise customers get the terms of their agreement.

7. Your data and your players' data

How the service handles message content, hashes, verdicts and player identifiers is described in the privacy notice at chatguard.dev/privacy, which forms part of these terms. In short: by default we store hashes and verdicts, not message text; evidence logging is a per-project opt-in with a retention you choose; player data can be erased by identifier. Where we process personal data on your behalf, we act as processor and you as controller, under the data processing agreement at chatguard.dev/dpa, which forms part of these terms on every plan.

You are responsible for telling your players that chat is moderated by an automated service and for having a lawful basis to send us their messages.

8. Intellectual property

Chat Guard, its model, word lists, documentation and dashboard remain our property or that of our licensors. You receive a non-exclusive, non-transferable right to use them through the API, the Unity package and the dashboard for as long as your organization exists and these terms are respected. The Unity package is published under the license in its repository. You keep all rights to your game and to the messages you send us; you grant us the right to process them to provide the service and to improve the local word lists in aggregate, never by publishing message text.

9. Warranties and liability

The service is provided as is. To the extent the law allows, we disclaim implied warranties, including fitness for a particular purpose, and we do not warrant that verdicts are accurate or that harmful content will always be caught or harmless content never flagged.

To the extent the law allows, our total liability for all claims arising out of the service in any 12-month period is limited to the fees you paid us for the service in that period, and we are not liable for indirect or consequential loss, lost profits or lost players. Nothing in these terms limits liability that cannot be limited by law.

10. Termination

You can delete projects and keys at any time, cancel a paid plan through the billing page, and ask us to delete your organization by email. We may terminate or suspend an organization for breach of these terms, for non-payment, or if we discontinue the service, with notice where practicable. After termination we delete organization data within 30 days, except what we must keep for accounting or legal reasons.

11. Changes to these terms

We may update these terms. Material changes are announced by email to organization owners or in the dashboard at least 14 days before they take effect; continued use after that date is acceptance. The current version is always at chatguard.dev/terms.

12. Governing law

These terms are governed by the law of the Republic of Serbia, and disputes are heard by the competent court in Belgrade, without prejudice to mandatory consumer protection rules that may apply to you elsewhere.

Contact

Kirill Tolmachev pr Beograd, Belgrade, Serbia. Email: support@chatguard.dev.