Privacy

Last updated: July 27, 2026.

T661.ai is a service operated by Pipaco Tech (“Pipaco Tech,” “T661,” “we,” or “us”), the British Columbia, Canada-based business responsible for personal information handled by this service. The Pipaco Tech Privacy Officer is accountable for this policy and can be reached at privacy@t661.ai.

What we collect and why

• We store repository evidence metadata, not source files. The GitHub App grants read access so we can collect repository, branch, and tag names; commit identifiers, messages, bodies, authorship (including author or co-author names and email strings recorded in Git), author and committer dates, file paths, and change statistics; and relevant pull-request titles, descriptions, authors, dates, branches, status, and URLs. File contents and diffs are never persisted. GitHub can include patch fragments in a commit-detail response; T661 discards those fragments immediately and does not send them to Claude.
• Commit messages can contain sensitive text. They are part of your evidence and are stored encrypted at rest in our database (Supabase/PostgreSQL) and shown only to members of your company workspace.
• AI processing. When you request generation, technical evidence needed for the draft is sent to Anthropic's Claude API. This can include commit and pull-request text, dates, selected file paths and statistics, project descriptions, and company-reviewed technical narratives. T661 removes direct email addresses and omits structured company, repository, contributor, hours, and sign-off fields where the drafting task does not need them. Free-form evidence can itself contain identifying or confidential text, so do not put secrets in commit messages or narratives. Under Anthropic's standard commercial API terms, inputs and outputs are not used to train models by default and are deleted from its backend within 30 days, subject to its stated safety, legal, service, and contractual exceptions. See Anthropic's training policy and retention policy.
• Information you upload. Supporting records such as timesheets, calendars, payroll records, contracts, invoices, design notes, and test results are stored in private object storage and included only in packages you authorize. Uploaded supporting records are not sent to Claude. Do not upload more personal information than the claim requires.
• People and access. We store account email, workspace membership, roles, invitations, reviewer comments, final confirmations, and audit events. Members can see workspace data; invited reviewers can read and comment unless their role changes.
• Product-use milestones. We record limited events such as completing setup, connecting repositories, producing a preview, reviewing a log, viewing an estimate, or starting checkout. These events contain identifiers and coarse counts—not repository names, commit text, narratives, personal names, email addresses, or claim dollar amounts—and are used to improve setup and reliability.
• Billing. Payment details go to Stripe; we never see card numbers. We retain Stripe customer/subscription identifiers and subscription state to provide and administer the service.

We use this information only to authenticate users, provide the evidence and claim workflow, secure and troubleshoot the service, communicate requested operational notices, administer billing, and meet legal obligations. We do not sell personal information or use claim evidence for advertising.

Consent and choices

You may withdraw consent for optional processing at any time by disconnecting the GitHub App, archiving a repository or supporting document, changing notification preferences, or emailing the Privacy Officer. Disconnecting GitHub stops future collection but does not silently destroy existing working papers. Claude processing is necessary to generate AI-assisted drafts; if you do not consent to that processing, do not request generation and contact us about deletion or export. Withdrawal does not affect processing already lawfully completed or records we must retain by law.

Retention and deletion

• Account and claim content: kept while the account is open so the customer's audit record remains available. After a verified account-closure or deletion request, we delete it from live systems within 30 days unless the customer asks us to retain it or a legal hold applies.
• Archived evidence: preserved with the claim audit record until account deletion; “archive” is recoverable and is not represented as erasure.
• Unclaimed GitHub installations: deleted after 7 days. Completed webhook processing records are deleted after 90 days.
• Product milestones and notification history: kept for up to 24 months.
• Backups: deleted live data may remain in encrypted, access-restricted backups until those backups age out, no later than 90 days after live deletion. A legal obligation, security investigation, billing record requirement, or active dispute may require longer retention; we will identify the reason when it applies.

Service providers and international processing

• Supabase provides database/auth/storage, Vercel hosts the app, GitHub supplies repository evidence, Anthropic processes curation inputs, Stripe processes billing, and Resend delivers email. Data may be processed outside Canada by these providers even when the primary database is configured in a Canadian region.

Safeguards and incidents

• Access is limited by workspace membership and database access rules; sensitive server credentials are not exposed to browsers. No system is risk-free. We investigate suspected unauthorized access and notify affected parties when required by applicable law.

Access, correction, deletion, and complaints

Email privacy@t661.ai from your account email and describe whether you want access, correction, deletion, consent withdrawal, or to make a complaint. We acknowledge requests within 7 business days, verify identity and authority, and normally respond within 30 days. We will explain any lawful refusal, extension, fee, or retention exception. You may also correct many company and project fields in Settings and export workspace or claim records from the product.

If our Privacy Officer does not resolve a concern, you may contact the Office of the Information and Privacy Commissioner for British Columbia or the Office of the Privacy Commissioner of Canada.

We may update this policy as the service or law changes. Material changes will be posted here with a new effective date and, when appropriate, communicated to account owners.