How Bantay decides what can appear in public search.

Core rules

• Public results are risk signals, not legal conclusions.
• Search stays open, but report intake is gated before moderation starts.
• Evidence files stay private even when a signal is reviewed.
• Human review is required before a public signal can appear.
• Reviewers do not publish identifiers unless they can confirm the identifier belongs to the reported counterparty.
• Corrections and appeal requests stay open to affected users.

Before a report can enter review

• Anyone can run a lookup without creating an account.
• Report intake currently uses Google sign-in to anchor the reporter flow.
• A bot check protects the submission flow before review starts.
• At least one proof file is required before moderators can consider publication.

Who decides

Current public release decisions are made by allowlisted human admins. The codebase includes an internal AI draft helper for summary writing, but AI does not approve, reject, or publish a report on its own.

What reviewers look for

• A clear identifier such as a mobile wallet number, bank account, Facebook profile, or marketplace handle.
• At least one proof file, such as a receipt or chat screenshot.
• Enough narrative detail to understand the unresolved transaction.
• Evidence that the identifier belongs to the reported counterparty rather than the victim, reporter, or an unclear screenshot participant.
• Signs that the submission is more than a duplicate, prank, or ordinary dispute.

What can appear in public search

• Reviewed signal counts, recency, incident type, and channel.
• Evidence counts and a short fact-first public summary.
• Correction and appeal paths for affected users.

Bantay does not publish uploaded evidence files, reporter identity, internal review notes, or admin audit records in public search.

When a report may be held or rejected

• No proof files are attached.
• The identifier is unclear, incomplete, or does not match the summary.
• The reviewer cannot distinguish the reported counterparty from the victim or reporter.
• The evidence suggests a normal customer-service dispute rather than an unresolved transaction signal.
• The report appears malicious, duplicated, or low-confidence.

Corrections and appeals

If a reviewed public signal is inaccurate, outdated, resolved, or linked to a recycled number, submit an appeal from the public result so moderators can re-check the reviewed signal.