Terms of Service. No legalese where we can avoid it.

Sigilix is operated by Sigilix, Inc.. These Terms govern access to and use of the Sigilix service, marketing site, and APIs.

Sigilix provides (a) automated, ensemble-based code review on pull requests in repositories where you have installed our GitHub App, including findings, severity ratings, and a verdict; (b) a coding agent available through our CLI and through issue-tracker integrations such as Linear; and (c) a research chat. The service is provided on a tiered basis (Free, Pro, Max, Ultra, Team, Enterprise) — see /pricing.

• ·You must be at least 18 and authorized to install GitHub Apps on the repositories you connect.
• ·You are responsible for the actions taken under your account, including those of your team members on shared seats.
• ·Notify us promptly at support@sigilix.ai if you suspect unauthorized access.
• ·Paid subscriptions are billed through Stripe, our payment processor. Monthly plans are billed in arrears; annual plans are billed up front for the seat base, with usage still metered transparently. Failure to pay results in suspension after the grace period documented in /pricing.
• ·Within an organization you can assign each member the tier that fits them — not everyone needs Max or Ultra — and allocate usage per seat. Mixed-tier workspaces are supported.
• ·You may cancel or downgrade at any time; refunds are evaluated case-by-case (see Privacy and Billing for the policy summary).

• ·Do not use Sigilix to review or operate on code that you do not have the right to share with us. You represent that you have all necessary rights to the code and data we process on your behalf.
• ·Do not attempt to extract our system prompts, model selection logic, or proprietary scoring weights via prompt injection or reverse engineering.
• ·Do not use Sigilix to generate, refine, or distribute malware, exploit chains, or content that violates GitHub's Acceptable Use Policies.
• ·Do not exceed the rate limits of your tier or attempt to circumvent them. Unlimited tiers (Enterprise) require a written agreement.
• ·Do not resell Sigilix as a service to third parties without an executed reseller agreement.

Sigilix findings, agent-generated code, and chat responses are best-effort signal generated by language models. They are advisory only — not legally binding security clearance, regulatory compliance certification, or guaranteed coverage of all defects.

Sigilix-managed models are the default models tuned and supported for Sigilix memory, retrieval, and review workflows. Optional CLI BYOK (Bring Your Own Key) changes which provider model you route a session to; it does not guarantee that the selected model will accept, retain, or follow Sigilix-provided memory, instructions, or retrieved context. Frontier models can reject, argue with, or override injected context because of their own alignment, safety, and prior-knowledge behavior. Those provider-level behaviors are outside Sigilix's control.

You retain final authority over what to merge. Sigilix never merges code automatically — every pull request the agent opens and every suggestion it makes is reviewed and merged by a human on your team. Rely on the output as one input among many: human code review, automated tests, security scanning, and your own judgement.

Where the model-learning setting is enabled (see the Privacy Policy for scopes and plan defaults), you grant Sigilix a limited license to retain and process the covered content — with secrets and PII redacted first — solely for the plan-specific learning purposes described in the Privacy Policy. Covered content includes retained, redacted review, CLI, and chat signals where those signals are retained under the Privacy Policy. On Free, Pro, Max, and Ultra, that content is used to improve Sigilix-managed models, user experience, review quality, and safety. On Team and Enterprise, learning is off by default; if your admin opts in, retained learning stays internal to your organization and is not used to train models outside your organization.

The grant is tied to the toggle: disabling learning ends new retention immediately. Deleting your account permanently deletes your personal data and any user-scope retained learning data tied to that account. If your organization requests deletion of its retained learning corpus, that corpus is permanently deleted from Sigilix learning systems.

• ·You retain all rights to the code you submit. Beyond the limited learning grant in Section 06 (where enabled), we claim no license to your code except what's necessary to run the service and post results back to your tools.
• ·Sigilix retains all rights to the service itself: the orchestration logic, prompts, scoring, model routing, and the marketing material on sigilix.ai.
• ·Findings and agent-authored changes posted to your repositories become part of your repository under your existing GitHub Terms; we make no separate claim on them.

We target high availability but do not yet publish a contractual SLA. Status updates and incident history are at status.sigilix.ai. Enterprise customers may negotiate a written SLA as part of their order form, and a Data Processing Agreement (DPA) is available on request via legal@sigilix.ai.

Read this section carefully.

The service is provided "as is" without warranty of any kind. To the fullest extent permitted by law, Sigilix, Inc. disclaims all warranties, express or implied, including merchantability, fitness for a particular purpose, and non-infringement.

Our aggregate liability for any claim related to the service is limited to the amount you paid for the service in the twelve months preceding the claim. We are not liable for indirect, consequential, or incidental damages, including lost profits or lost data, even if advised of the possibility.

You may stop using Sigilix at any time by uninstalling the GitHub App, disconnecting any integrations, and canceling your subscription. We may suspend or terminate accounts for material breach of these terms, unpaid invoices, or activity that puts our infrastructure or other customers at risk. We will give reasonable notice for non-emergency terminations. On termination you may request deletion of your data, including any retained learning corpus, per the Privacy Policy.

These Terms are governed by the laws of [state — confirm with counsel], USA, without regard to conflict-of-laws provisions. Disputes will be resolved in the state or federal courts located in [venue — confirm with counsel], except where mandated otherwise by your local jurisdiction.

We will revise these Terms as the service evolves. Material changes will be announced by email to the billing contact and posted at sigilix.ai/terms with the new effective date. Continued use after a change constitutes acceptance.

Legal: legal@sigilix.ai. Privacy: privacy@sigilix.ai.