Terms of Service

Neolith is operated by Neolith, an adaptic.ai company ("Neolith", "we", "us"). These Terms form a binding agreement between you and Neolith. If you are entering into them for an organization, you represent that you are authorized to bind that organization. You must be able to form a legally binding contract to use the service.

The service is provided as a pre-release beta. It is offered for evaluation and collaboration with a limited cohort of design partners. The beta:

• may contain defects, change without notice, and be modified, suspended, or discontinued at any time;
• is offered without charge for the service itself for the duration of your beta engagement (see §07 on compute);
• is access-controlled — we admit organizations by fit, in waves, and may decline or revoke access at our discretion.

You should not rely on the beta for any use where failure, downtime, or error could cause material harm without your own independent controls in place.

A workspace is your organization's instance of Neolith. Members authenticate through your identity provider, and roles are governed by your workspace's access controls (RBAC). You are responsible for:

• the security of your credentials, identity provider, and any secrets you store in the workspace;
• the acts of your members and the human owners you designate for decisions and approvals;
• configuring charters, policies, and approval gates appropriately for your risk posture.

Neolith lets you operate AI colleagues — software agents that plan, draft, execute, and propose decisions alongside human members, each operating under a charter you define. You understand and agree that:

• You remain responsible. AI colleagues act on your behalf and within your workspace. You are responsible for the charters, guardrails, and approval gates you configure, and for the outputs and actions they produce.
• Autonomy is bounded by your configuration. The service provides controls — scoped charters, policy thresholds, propose-and-approve flows, and human-gated release for irreversible acts — but it is your configuration that determines what an agent may do alone versus what routes to a named human.
• Humans hold the agenda. Consequential, irreversible, or externally-facing actions should be gated on a human owner. You agree not to disable controls in a way that lets an agent take irreversible external action without human authorization where the law or your obligations require one.
• Outputs may be imperfect. AI-generated content can be inaccurate, incomplete, or unsuitable. You are responsible for reviewing outputs before relying on or publishing them.

You agree not to use the service, or permit any member or AI colleague to use it, to:

• break the law, infringe others' rights, or violate the terms of any connected provider or model;
• generate or distribute unlawful, harmful, deceptive, or infringing content;
• circumvent access controls, probe or disrupt the service's security or integrity, or reverse-engineer it except as permitted by law;
• take autonomous external actions (trades, transfers, publications, or legal commitments) without the human authorization your obligations require.

You retain all rights in the content, documents, messages, and data your members and AI colleagues create or upload ("Customer Content"). You grant us a limited license to host, process, and transmit Customer Content solely to provide and improve the service, and to route it to the model providers and subprocessors needed to run your AI colleagues. We do not sell Customer Content, and we do not use it to train foundation models. Our handling of personal data is described in the Privacy Policy.

In Neolith, compute is payroll: an AI colleague's cost is its machine and its model time. During the beta:

• the service is provided at no charge for your beta engagement;
• if you run colleagues on managed compute we provide, that compute is included within fair-use limits we may communicate to you;
• if you bring your own provider keys or machines, you bear the underlying compute cost directly with that provider; we surface those figures in Fleet but do not mark them up.

We will give design partners clear, advance notice and preferential terms before any part of the service becomes paid.

The service records events — messages, decisions, approvals, and system actions — to a single audit timeline (the "decision ledger") to make your company's activity attributable and replayable. You control your workspace's records subject to these Terms and the Privacy Policy. Some records are intended to be append-only so the audit trail stays trustworthy; where you require deletion, we will honor lawful requests in a way that preserves the integrity of the remaining trail.

The beta and any non-public information we share (features, roadmap, performance) are confidential. You may share feedback, and you grant us a perpetual, royalty-free license to use that feedback to improve the service without obligation to you. Please don't publicly benchmark or disclose non-public aspects of the beta without our written consent.

We aim for a dependable service but make no uptime commitment during the beta. We may add, change, or remove features, and we may update these Terms. If we make a material change, we'll take reasonable steps to notify you; continued use after a change means you accept it.

The service is provided "as is" and "as available", without warranties of any kind, whether express, implied, or statutory, including merchantability, fitness for a particular purpose, and non-infringement. We do not warrant that the service will be uninterrupted, error-free, or that AI-generated outputs will be accurate or suitable for your purposes.

To the maximum extent permitted by law, neither party is liable for indirect, incidental, special, consequential, or punitive damages, or for lost profits, revenue, data, or goodwill. Because the beta is provided without charge for the service, our aggregate liability arising out of or relating to the service is limited to the greater of the fees you paid us for the service in the three months before the claim (which may be zero) or US$100.

These Terms apply while you have access to the beta. Either party may end the engagement at any time. On termination, your right to use the service stops; we will make your Customer Content available for export for a reasonable period and then delete it in the ordinary course, except where retention is required by law. Sections that by their nature should survive (content rights, confidentiality, disclaimers, liability, governing law) survive termination.

These Terms are governed by the laws of the Dubai International Financial Centre (DIFC), without regard to conflict-of-laws rules, and the DIFC Courts have exclusive jurisdiction — unless your signed design-partner agreement specifies otherwise, in which case that agreement controls. If any provision is unenforceable, the rest remains in effect.