14 promises, structured.

violates

Any commercial transfer of user identifiers, content, behavioural signals, or derived profiles to a third party for compensation. Includes anonymised-but-rejoinable datasets.

burden of compliance

operator

audit mechanism

Quarterly third-party audit of data-egress logs. Charter members may request raw audit reports.

escalation path

Board-level escalation. Charter consultation required before any change. Pact-violation form on /pact triggers founder email.

violates

Any restriction on export, any failure to honour a delete request beyond the documented 30-day technical-backup window, any retention of derived data after deletion.

burden of compliance

operator

audit mechanism

End-to-end test of export + delete on every release. SLA: export within 24h, delete effective immediately.

escalation path

Failure to honour deletion = Pact violation, founder-direct response.

violates

Any default that broadens scope without opt-in. Any silent re-scoping of past content. Any UI that obscures the current scope.

burden of compliance

platform-content

audit mechanism

Compose flow shows current scope above the fold. Past whispers retain their original scope through any UI redesign.

escalation path

Charter review of any compose-flow scope change.

violates

Any third-party analytics that fingerprints users. Any ad-tech or marketing pixel. Any cross-site identifier propagation.

burden of compliance

operator

audit mechanism

External CSP audit. No third-party scripts on chatter.today domains. Headers-only third-party calls.

escalation path

Board-level. New third-party integrations require Charter sign-off.

violates

Use of phone/email verification for any purpose other than abuse prevention and account recovery. Retention of verification artifacts beyond what's needed.

burden of compliance

operator

audit mechanism

Twilio Verify logs reviewed quarterly. Verification artifacts purged within 30 days.

escalation path

Founder-direct review.

violates

Personal attacks unrelated to ideas, dogpiles designed to harm, content that targets a person's character rather than their argument.

burden of compliance

moderation

audit mechanism

Moderation queue at /moderation. Public moderation log (anonymised).

escalation path

Charter members may flag dogpiles directly to founder.

violates

Same as V1, refined: focus is on the target. Disagreement with someone's reasoning ≠ disagreement with their existence.

burden of compliance

moderation

audit mechanism

Same as V1.

escalation path

Same as V1.

violates

Any content targeting people for hate based on race, religion, gender, sexuality, disability, or immigration status. No grey area, no warnings, no second chances.

burden of compliance

moderation

audit mechanism

Permanent-ban actions logged with original content + decision rationale, accessible to charter on request.

escalation path

Immediate action; appeal window does NOT apply to V3 violations.

violates

Doxxing, threats, sexual harassment, sustained unwanted contact, content that endangers a real person's safety.

burden of compliance

moderation

audit mechanism

Reports → moderation queue → action within 72h. Doxxing/threats acted on immediately.

escalation path

Founder-direct for severe cases.

violates

Engagement-bait, rage-bait, and outrage cycles whose primary value is generating reaction rather than informing.

burden of compliance

platform-content

audit mechanism

Heat-score algorithm reviewed for outrage-amplification bias.

escalation path

Charter review of feed-ranking changes.

violates

Any inclusion of user-generated content in datasets used to train third-party or external AI models. Includes derivative datasets.

burden of compliance

operator

audit mechanism

Annual third-party data-flow audit specifically for AI training inclusion. Public attestation.

escalation path

Permanent commitment, non-modifiable carve-out in Terms §4.

violates

Any AI feature that modifies feed ranking, suggested posts, or composed text without explicit per-user opt-in. Conversely, disabling safety AI is a V3-class violation.

burden of compliance

platform-content

audit mechanism

Settings → AI shows every AI feature, on/off state, what it does. New AI features require new opt-in.

escalation path

Charter consultation before any new AI feature ships.

violates

Any auto-generation of content attributed to your account. Includes 'reply suggestions' that auto-send.

burden of compliance

platform-content

audit mechanism

Compose surfaces are explicit-submit only. No background-AI-write features.

escalation path

Permanent commitment.

violates

Any UI surface that exposes a user's follower count, like count, view count, or reach metric to anyone other than the user themselves.

burden of compliance

platform-content

audit mechanism

Visual audit on every UI release. Charter members may flag any popularity-score surface.

escalation path

Charter review. Vouches received ARE visible — credentials, not popularity.