We are finalising the Tailor terms of service. This page is a work-in-progress placeholder — no clauses, waivers, or obligations have been finalised yet. The points below describe the contact path for terms enquiries today. This page will be replaced with the full terms once they are ready.
- Contact legal@tailor.au for any terms-of-service enquiries, including requests for a written agreement ahead of the finalised terms.
- Customer agreements currently in place (e.g. signed order forms, MSAs, or contract schedules) remain the governing documents between Tailor Intelligence Pty Ltd and those customers.
- Nothing on this page constitutes final legal terms.
If you are a Tailor customer and need written confirmation of the terms that apply to your use of the service for your own compliance reviews, email legal@tailor.au and we will respond directly.
AI agent connectors
When you authorise a third-party AI runtime (for example ChatGPT through OpenAI Apps SDK or Codex, Claude through Anthropic remote MCP or the Claude Code plugin marketplace, Cursor, Windsurf) to connect to Tailor via tailor.au/connect, the following clauses apply in addition to any signed agreement.
- Tenant boundary. Every connector client and every issued token is bound to a single tenant. Tailor enforces tenant scope on every tool call; cross-tenant access is refused and audit-logged.
- Least-privilege scopes. You authorise specific scopes (read, write, destructive) at consent time. Tailor honours the granted scope on every tool invocation regardless of what the agent requests at runtime.
- Revocation. You may revoke any connector client or token at any time. Revocation propagates within seconds and invalidates outstanding tokens immediately.
- AI runtime relationship. Tailor is not affiliated with, endorsed by, or a reseller of the third-party AI runtime. The AI runtime vendor's own terms of service govern your use of that runtime; Tailor's terms govern only the Tailor side of the connection.
- No raw secret echo. Tailor will never return your API keys, OAuth codes, refresh tokens, or other shared secrets in a tool response or audit log line. Outputs are passed through a sensitive-value redactor so the agent's context window and any downstream training surface cannot be used to reconstruct your credentials.