Content Moderation API
OmniMod moderates user content with auditable decisions, policies you can set and tune, and actions your product can enforce.
Decision intelligenceMulti-lingualVersioned policies by modalityAudit-ready outcomesCustom moderation requests
Result214ms · cache miss
Submitted textNice work, genius. Keep this up and see what happens.
Recommended actionHold for review
ReasonContextual harassment
ConfidenceMedium
Ambiguous intent. Policy recommends review instead of automatic rejection.
community-safety v7rules + semantic review
What we moderate
OmniMod reviews text and images, applies your policy posture, and returns an auditable decision your product can enforce.
TextText policy
Detect harassment, hate, spam, threats, sexual content, self-harm signals, URL evasion, and mixed-language abuse.
Language coverageEnglish, Spanish, French, and German rule packs, with semantic review for mixed or unknown language cases when enabled.
Tunable thresholdsLanguage-aware routingSemantic review
ImageImage policy
Detect visual safety signals in images up to 4 MB, then apply your image policy to decide what should pass, flag, or review.
Image coverageImage findings become policy-aware outcomes with evidence your team can inspect later.
Scored evidenceTunable thresholdsReview context
DecisionActionable output
Return the moderation outcome, recommended action, confidence, explanation, policy version, and decision intelligence for apps and agents.
Audit contextEach result carries enough context for product enforcement, support review, and policy tuning.
AllowFlagBlockReviewPolicy versionReason codeConfidenceRecommended action
Custom Policies
Start with OmniMod defaults, then tune text and image policies separately as your product matures.
Set different flag and block levels by modality and category.
Old decisions keep their original policy context after tuning.
Your policies, keys, usage, and cached decisions stay scoped to your tenant.
Platform
OmniMod turns moderation decisions into controls your team can inspect, tune, and enforce.
Event inspectorCustomer dashboard
Policy, routing, language, cache, and analyzer evidence stay attached to the event.
flag
Policycommunity-safety v7
Routingrules + semantic
Languagemixed · 0.84
12:41 PMcontextual harassmentflagmiss
12:38 PMclean community replyallowhit
12:33 PMimage policy thresholdflagmiss
Tune text and image thresholds independently without losing the policy context behind older decisions.
Text policyv7 active
Image policyv3 active
Track plan limits before moderation pauses and keep subscription state visible to operators.
Monthly moderation units640 / 1,100 used
KeysNamed API keys with revoke and rotate flows.
EvidenceAnalyzer scores and reason codes for audit review.
PrivacyHashes and metadata by default, not raw text storage.
OperationsEmail alerts for quota and account lifecycle events.
Pricing
One text unit is up to 1,000 characters. One image unit is one image up to 4 MB.
Free trial
$0one-time beta trialStarter trial credits for approved beta users
Founder beta
Launch
$9per month$7/mo for first 2 months with Founder promo1,000 text units + 100 image units / month
Scale
$29per month5,000 text units + 500 image units / month
Growth
$79per month20,000 text units + 2,000 image units / month
Custom
Customcontact usCustom limits, volume pricing, and fit review
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。
