Multi-agent DAG orchestration, purpose-built for enterprise engineering teams Learn more →

Safety

Founders build. Users do unexpected things.

Cendriix bakes user-content safety into every app it builds, abuse reports, content moderation, rate-limiting, age-gating, and incident response. You don't configure it. It is there when the first user arrives. Everything stays inside Cendriix, reports, moderation logs, escalation history, and law-enforcement cooperation records.

What Cendriix bakes into every app it builds

These are not optional add-ons. They are part of the default build. You can tune thresholds; you cannot remove them from apps with user-generated content.

Abuse report endpoint

Auto-wired to every app Cendriix builds. Reports route to the workspace owner with severity triage.

Content moderation

Text and image moderation on user-generated content. Video moderation available on Enterprise. Tunable thresholds per workspace.

Rate-limiting with adaptive backoff

Progressive slowdown before hard blocks. Bot traffic is shaped separately from human sessions.

Bot and spam detection

Behavioural fingerprinting, headless-browser detection, and reputation scoring on new accounts.

Age gates

Deployed automatically for apps serving alcohol, gambling, adult content, or other age-restricted categories.

Geographic blocks

Sanction-list enforcement (OFAC, EU, UN) and age-of-consent jurisdiction controls applied at the network edge.

Right-to-erasure handler

"Delete me" requests are queued, acknowledged within 24 hours, and completed within 30 days. GDPR Art. 17 SLA tracked per request.

Subject Access Request handler

Data export requests fulfilled automatically with operator review before release. DSAR log maintained for regulatory audit.

Account takeover detection

Geo-velocity, device-fingerprint, and credential-breach signals trigger step-up auth without requiring the user to do anything first.

2FA enrollment prompts

High-value actions (payment, export, admin role grant) trigger a 2FA prompt if the session does not already have a strong factor.

Moderator action audit trail

Every content moderation decision, allow, remove, escalate, is logged with actor, timestamp, and reason. Immutable.

Banned-user re-attempt protection

Device, email, and payment fingerprinting detect banned users re-registering and apply the same restrictions automatically.

What happens when abuse is reported

Six steps from report to resolution. All tracked inside Cendriix.

1
Report received
User submits a report via the in-app endpoint. A ticket is created in Cendriix and the workspace owner is notified immediately.
0 min
2
Automatic severity triage
Cendriix classifies the report: spam, CSAM, violence, harassment, or other. CSAM triggers an immediate escalation to Cendriix's trust-and-safety team, bypassing the workspace owner review step.
< 2 min
3
Reported content quarantined
For high-severity reports (CSAM, violence, credible threat), the reported content is immediately hidden pending review. The reported user's ability to create new content may be restricted.
< 5 min
4
Workspace owner reviews
The workspace owner receives an in-Cendriix task with the report, the reported content, the user's history, and a recommended action. Three choices: dismiss, warn, suspend.
< 24 h
5
Action taken and logged
The workspace owner's decision is applied and logged. The reporting user receives a confirmation. The outcome is appended to the trust-and-safety audit trail.
Same session
6
Pattern detection
Cendriix analyses the reported content against patterns across the workspace. If the report reveals a systematic issue, a workspace-wide review is triggered automatically.
< 48 h

Severe-incident protocol

What Cendriix does if an app it built is used for serious harm, illegal content, large-scale fraud, or organised abuse.

01Detection

Cendriix monitors for signals: CSAM hash matches (PhotoDNA integration), fraud ring fingerprints, regulatory complaints, and law-enforcement preservation requests.

02Triage

Severity classification by the trust-and-safety team within 1 hour. The workspace owner is notified. The affected content is quarantined pending outcome.

03Action

Graduated response: warn the workspace owner → throttle the offending traffic → suspend affected user accounts → suspend the workspace → terminate the workspace and revoke API access. Each step is documented.

04Law-enforcement cooperation

Cendriix cooperates with lawful requests under proper legal process only, a valid court order, subpoena, or equivalent. We do not respond to informal requests. We notify affected workspace owners unless legally prohibited from doing so.

Trust and safety contacts

Every address goes to a human team, not a bot, not a ticketing queue that silently expires after 30 days.

AddressPurposeSLA
abuse@cendriix.aiGeneral content abuse reports (spam, harassment, fraud)< 24 h
csam@cendriix.aiCSAM, escalated immediately, 24/7 on-call response< 1 h
lawenforcement@cendriix.aiLawful-process requests (subpoenas, court orders, NatSec letters)Per legal timeline
security@cendriix.aiVulnerability disclosures, coordinated security research< 48 h

For law-enforcement requests: Cendriix requires a valid legal process (subpoena, court order, or national-security letter) before disclosing any user data. We do not respond to informal requests from law enforcement or third parties. We maintain a transparency report of all legal requests received, published semi-annually.

Questions about content safety?

Trust-and-safety questions, platform policy concerns, and abuse escalations are all handled by a team that reads every message.