Terms, privacy and AI notice

Use at your own risk and limitation of liability

Use Lexeze at your own risk. The tool and its output are provided as is and as available, without warranties of accuracy, completeness, fitness for a particular purpose or uninterrupted availability. To the fullest extent permitted by applicable law, Lexeze and its operator are not liable for any loss, damage, cost or claim arising from your use of, inability to use, or reliance on the tool or its output. Nothing in this notice excludes or limits liability or statutory rights that cannot lawfully be excluded or limited.

This is a pre-production tool for enrolled advocates and their authorized staff. Compliance and operational approvals remain incomplete. Use fictional or anonymized test matters only; do not submit live client data or rely on output for client advice or filings. An enrolled advocate must independently verify all facts, citations and applicable law. This acknowledgement does not resolve outstanding compliance requirements or authorize a paid production launch.

Operator and contact

Operator identity pending - not approved for production use.

Registered address pending.

Grievances, privacy requests and errors: Grievance contact pending - do not submit real client data.

Terms and acceptable use

Lexeze supplies AI-assisted drafting tools and general information. It does not provide legal advice, practise as an advocate, certify title or create an advocate-client relationship. An enrolled advocate must independently verify every relevant fact, source document, search, finding and applicable law before signing, filing or supplying an opinion. Staff must refer output to their supervising advocate.

Do not impersonate counsel, claim guaranteed outcomes, submit unlawfully obtained data, infringe third-party rights, bypass access controls, or use unreviewed output as final advice. You must have authority and an applicable lawful basis to process client and third-party information. Do not submit unnecessary Aadhaar, payment-card, health or other sensitive identifiers, or children's personal data. This service does not support child-data consent verification.

Generated drafts may contain hallucinated facts, omissions, obsolete law or errors in state-specific rules. Source checks are aids, not proof of correctness. There is no guarantee of accuracy, legal effect, or acceptance by a bank or court. Statutory consumer rights remain unaffected. These terms do not impose a liability cap or mandatory arbitration. Operator-specific commercial terms require legal review before a paid launch.

What data is processed and why

Account identifiers and password hashes support sign-in. Google sign-in uses your profile and email. Saved report contents are stored in server-side private files; report titles, account identifiers, client names, contact details and notes are stored in databases. Consent timestamps and report save/review/delete events support accountability. These records can contain personal information even where they do not contain full matter text.

The report studio uses Google Gemini for notes, automatic prefill and findings suggestions. Form drafting can use the selected available provider: Google Gemini, Anthropic Claude or OpenAI. The submitted context is sent to that provider to prepare drafts. This can involve processing outside India. Provider regions, retention and contractual protections must be approved by the operator before live use; selecting a model is not evidence of a no-training agreement. Lexeze has no customer-data training or fine-tuning feature.

Unsubmitted draft state stays in the browser tab, but background prefill transmits entered notes after consent. Saved matters remain until deletion or the operator's documented retention process. Security records and backups have separate retention and legal-hold requirements. The operator must publish approved periods and vendor details before production activation.

Approved active-data retention: reports, not yet configured; unlinked client records, not yet configured. Erasure is subject to legal holds and the operator running the retention process. Linked client records are retained while their reports remain.

In configured production deployments, report contents, report titles, client fields and rights-request descriptions use application-level encryption. Identifiers, timestamps and authentication/consent metadata still require encrypted storage volumes and protected backups. Audit events contain pseudonymous actors and record fingerprints, not raw prompts. Legal holds can prevent changes or erasure; withdrawn consent still stops new matter processing.

Your controls

Terms, AI processing and privacy processing require separate affirmative choices before matter processing. Withdraw consent from Privacy controls to stop new processing; withdrawing does not retroactively erase data. Export and erase your active workspace there, including after withdrawal. Reports can be corrected and deleted in the existing workspace. Contact the grievance officer for account deletion, correction, access, nomination, grievances, provider deletion and backup handling, or to report inaccurate output. The operator must verify identity and communicate applicable deadlines and any lawful retention exceptions.

Privacy controls also provides a tracked request form for access, corrections, erasure, nomination, account deletion, grievances and output errors. It remains available after consent withdrawal. Check the status and resolution there; the app does not send email notifications. The operator must independently verify authority for requests concerning third parties.