Terms of Service

These terms govern your use of Trellaris (the “Service”): the web application at trellaris.app, its API at api.trellaris.app, the documentation, and any agent, automation or integration you run through them. “We” and “us” means the operator of the Service. “You” means the person signing in and, where you act for an organisation, that organisation.

By creating an account, signing in with a social or company identity, or using the demo organisation, you accept these terms and the Privacy Policy. If you accept on behalf of an organisation, you confirm you are authorised to bind it. If an organisation has a separate written agreement with us, that agreement prevails where the two differ.

You sign in through an identity provider we support (Google, GitHub, or your company’s single sign-on). We do not store a password for you. You are responsible for the security of that identity and for everything done under it until you tell us it was compromised.

Every account belongs to an organisation. Depending on how your organisation is configured you land in a personal organisation of your own, in the organisation that has claimed your email domain, or in a shared landing organisation. An organisation administrator can add and remove members, assign roles, set policies, place legal holds and stop every agent in the organisation. Those actions are theirs, not ours.

You must be at least 18 years old and able to enter a contract to use the Service.

The one-click demo accounts open a single organisation shared with every visitor. Anything you type, upload or build there can be seen, changed or deleted by anyone else using the demo, and we may reset it at any time without notice.

Do not put confidential, personal or regulated data into the demo organisation. Sign in with your own identity to keep your work to yourself.

You agree not to use the Service, or any agent you build on it, to:

• break the law, or help anyone else to;
• access data, systems or accounts you are not authorised to access, including through a connector or an agent acting for you;
• send spam, malware or unsolicited messages, or run automations against third parties who have not agreed to receive them;
• attempt to defeat the platform’s controls — data-loss prevention, approvals, allow-lists, budgets, sandbox isolation, rate limits or the kill switches;
• probe, scan or load-test the Service beyond what your organisation has agreed with us in writing;
• generate content that is defamatory, harassing, sexually explicit involving minors, or that infringes another person’s rights;
• resell, sublicense or offer the Service to third parties as your own product without our written agreement.

Agent definitions, skills, tool servers and connectors you publish to a shared ring pass automated scans before other people can use them. A scan is a control, not a warranty: you remain responsible for what you publish.

You keep ownership of everything you bring to the Service: your messages, documents, knowledge collections, memories, agent and skill definitions, policies, and the outputs your agents produce for you (together, “Your Content”).

You grant us the limited licence we need to run the Service for you: to store Your Content, pass it to the AI model providers and connectors you have configured, index it for search, show it to the people your organisation’s visibility rules allow, and keep the copies our backups and audit obligations require. We do not use Your Content to train models, and we do not sell it.

You are responsible for having the rights to whatever you upload and for the instructions you give an agent. Where an agent acts on your behalf through a connector, it does so with your credentials and under your authorisation.

Agents on the Service are powered by large language models. Their output can be wrong, incomplete, out of date or biased, however confident it sounds. Cited answers cite the document the agent read; they do not guarantee the document is correct.

Output is information, not advice. Nothing an agent says is legal, financial, medical, tax or investment advice, and the built-in market watcher in particular describes markets and never recommends a trade. Check anything that matters before you rely on it, and keep a human in the loop for decisions with consequences.

The Service asks a person to approve an agent’s writes by default. If your organisation raises an agent’s autonomy so it can act without that approval, the organisation accepts the result.

The Service connects to services you choose: AI model providers, Google Workspace, ServiceNow, Slack, Microsoft Teams, MCP tool servers, remote agents and others. Each has its own terms, and by connecting one you agree to them. We are not responsible for what a third-party service does with a request an agent sends it, or for its availability.

Connecting a service stores a credential for you, encrypted, which you can revoke at any time from Settings. Revocation takes effect on the next call.

The demo organisation and self-serve sign-in are currently provided without charge. Model usage is metered and shown to your organisation as cost per run; where a paid plan applies, the price, billing period and usage limits are those in your order form or the plan you selected, and taxes are added where required.

We may introduce or change fees for a plan with at least 30 days’ notice. Continuing to use the plan after that date is acceptance of the new fee.

The Service, its software, design, documentation, built-in agents, skills and templates, and the Trellaris name and mark are ours or our licensors’. These terms give you a non-exclusive, non-transferable right to use them through the Service and nothing more. Feedback you give us about the Service may be used without obligation to you.

You can stop using the Service at any time; an organisation administrator can remove members and close the organisation. We may suspend or end access, with notice where practical, if you breach these terms, if your use creates a security, legal or operational risk, or if the Service is discontinued.

When access ends, Your Content is deleted on the schedule described in the Privacy Policy, except where a legal hold or a legal obligation requires us to keep it. Sections on your content, AI output, disclaimers, liability and disputes survive termination.

The Service is provided “as is” and “as available”. To the fullest extent the law allows, we disclaim all warranties, express or implied, including merchantability, fitness for a particular purpose, non-infringement and any warranty that the Service or an agent’s output will be accurate, uninterrupted or error-free. The demo organisation and any feature marked preview, beta or experimental may change or disappear without notice.

To the fullest extent the law allows, neither we nor our suppliers are liable for any indirect, incidental, special, consequential or punitive damages, or for lost profits, revenue, data or goodwill, arising from the Service or from anything an agent did or failed to do, even if we were told such damages were possible.

Our total liability for all claims arising from the Service in any twelve-month period is limited to the amount you paid us for the Service in that period, or one hundred US dollars if you paid nothing. Nothing in these terms limits liability that cannot be limited by law.

You will defend and indemnify us against claims, losses and costs arising from Your Content, from your breach of these terms, or from an agent acting on instructions or credentials you supplied, except to the extent we caused the loss.

We improve the Service continuously and may add, change or retire features. We may also update these terms. For material changes we will give notice in the application or by email at least 14 days before they take effect; using the Service after that date is acceptance. The effective date at the top of this page is the current version.

These terms are governed by the laws of the jurisdiction in which the operator of the Service is established, without regard to its conflict-of-law rules, and the courts of that jurisdiction have exclusive jurisdiction, unless your organisation’s written agreement with us says otherwise. Before starting a claim, either side will write to the other and allow 30 days to resolve it informally.

Questions about these terms go to legal@trellaris.app. Security reports are welcome at the same address and are acknowledged within two working days.