Trust & Safety

Speakeazy is a location-based social app where every venue chatroom is a room you are physically standing in, with real people whose identities we've verified. This page describes the concrete steps we take to prevent, detect, and act on abusive content and behavior. It exists so our users, our partners, and the platforms we distribute through can audit our process without asking us for details in private.

1. Prevention

Most abuse is prevented before it happens by controlling who is in the room in the first place.

• Real-name identity. Users sign in via Apple, Google, or email magic link. Display name and profile photo default to the identity provider's values. Aliases intended to hide identity violate our Community Guidelines.
• Phone verification. Every account verifies a phone number via one-time SMS (Twilio). This is a friction gate against throwaway accounts.
• Government-ID verification. For age-restricted (21+) venues such as bars, users must complete Persona identity verification, which matches a government-issued ID against a live selfie. Users who fail verification are blocked from posting to those venues.
• Geofence enforcement. A user can only post in a venue chatroom while physically inside the venue's 100-meter geofence. Location spoofing is a bannable violation.
• Ephemeral chat. Venue chat messages auto-delete 24 hours after posting. This limits the shelf-life of harmful content and shapes user expectations toward the present moment.

2. Detection

We rely primarily on the community and on human review, with automated signals as a supplement.

• User reporting. Every user, message, video, and direct-message thread has a one-tap Report action. Reports specify a target type (user, message, video, or private chat), a target ID, and a free-text reason (1–500 characters).
• Duplicate suppression. Reports are deduplicated per (reporter, target) pair within a rolling 60-minute window, so a spam loop cannot flood the moderation queue or crowd out legitimate reports. The client receives an identical success response either way, so an abusive user cannot detect throttling.
• Blocking. Any user can block any other user. Blocks are reciprocal: the blocked user is hidden from the blocker's rosters and cannot view the blocker's profile. Blocking is a normal, healthy use of the app and is encouraged in our Community Guidelines.
• Automated content signals. We run pre-publish screening on chat messages and images. This is a supplement to human review, not a replacement — every report goes to a human moderator regardless of what any automated classifier said.

3. Response

Every report enters the same queue, is reviewed by a human moderator, and results in one of the actions below.

• Target response time: 24 hours. This is the same commitment published in our Terms of Service and Community Guidelines.
• Resolution states. Every report is closed with one of three explicit resolutions: dismissed (does not violate Guidelines), warned (content removed, user notified), or banned (account terminated).
• Cascade actions. When resolving a report against a user, a moderator can atomically ban the target account in the same action. The report resolution and the ban are recorded as a single moderator decision.
• Audit trail. Every moderator action — resolving a report, banning a user, restoring a user, adjusting verification — is written to an append-only audit log with the moderator's identity, the before/after state of the affected record, and the reason provided. The audit log is not user-editable.

4. Enforcement tiers

We match the severity of the action to the severity of the violation.

• No action. The content or behavior does not violate our Guidelines.
• Content removal. The specific message, photo, or profile element is deleted. Used for isolated Guidelines violations from users with no relevant history.
• Warning. The offending content is removed and the user is notified in-app. Used for first-time minor violations.
• Temporary suspension. The account is locked for a defined period. Used for repeat or moderate violations.
• Permanent ban. The account is terminated. Used for severe violations (harassment campaigns, threats, sexual content involving minors, coordinated fraud), for repeated offenses across tiers, and for any attempt to evade a prior suspension.

Bans are enforced at the account level and, where necessary, at the device level. Attempting to create a new account after a ban is itself a bannable violation.

5. Zero-tolerance categories

Some categories skip the escalation tiers and result in an immediate permanent ban on first substantiated report.

• Content that sexualizes minors. Referred to the National Center for Missing & Exploited Children (NCMEC) and law enforcement where applicable.
• Non-consensual intimate imagery.
• Credible threats of violence.
• Coordinated fraud, phishing, or scams targeting the app's users.

6. Appeals

Users can appeal any moderation action by emailing appeals@speakeazysocial.com with the affected account's email and a short explanation. Appeals are reviewed by a moderator who did not participate in the original decision. Our target response time for appeals is 5 business days.

7. Transparency and contact

Reports of urgent trust & safety incidents: safety@speakeazysocial.com. Law-enforcement inquiries: privacy@speakeazysocial.com. General questions about our moderation process: hello@speakeazysocial.com.

This document will be revised as our tooling and process evolve. The Last updated date at the top reflects the most recent substantive change.