SafetyKit
SafetyKit is an AI-native trust, safety, and compliance platform that replaces human risk reviewers with autonomous agents, serving Fortune 500 marketplaces, payment platforms, card networks, and AI platforms with multi-modal content moderation, fraud prevention, and merchant intelligence at scale.
- Company typePrivate
- Founded2022
- HeadquartersSan Francisco, United States
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What SafetyKit does
SafetyKit is an AI-native trust, safety, and compliance platform that replaces human risk reviewers with autonomous AI agents. The platform ingests text, images, video, audio, user events, transactions, and URLs at 4,200 entities per second, processing 20 billion LLM tokens daily across its Fortune 500 customer base. Its core engine analyzes entire entities in context—flagging fraud rings, reused data elements, and coordinated violations across accounts and devices—and returns enforcement actions with confidence scores at sub-3,200ms latency. Reported performance metrics include 95%+ accuracy versus an 80% human reviewer baseline and a 94% fraud detection rate with under 0.1% false positives.
The product suite comprises four primary solutions—Content Moderation with AI, Fraud Prevention with AI, Merchant Intelligence, and Networks (network-scale takedowns)—plus modular add-ons including Credit Risk, Merchant Monitoring, AI Ops (autonomous agent workflows), ML Ops (custom classifiers, clustering, anomaly detection), Human Ops (case management), Compliance Solutions covering EU DSA, UK OSA, and the Take It Down Act, and a Policy Library of 200+ pre-built policies. The platform is built on a serverless AWS stack (Lambda, DynamoDB, S3, CDK) managed within a TypeScript monorepo and is exposed via REST API and SDKs in TypeScript, Python, Kotlin, and Java. AI agents perform end-to-end investigations, pulling context from multiple sources, verifying documents, and mapping fraud networks before taking action or escalating edge cases to human reviewers.
SafetyKit operates an API-first, enterprise sales motion targeting online marketplaces, payment platforms, card networks, financial institutions, and LLM platforms. Revenue is generated through SaaS/API subscriptions on quote-based enterprise contracts, with pricing tied to volume and usage. The company is headquartered in San Francisco, employs 11–50 people (12 engineers), and serves customers including Upwork, Eventbrite, Etsy, Block, Square, Faire, Patreon, Kickstarter, Lime, Character.ai, and a major card network processing over $700B in transactions. It raised a $27M seed round in April 2025 from Ribbit Capital, First Round Capital, and Y Combinator.
SafetyKit firmographics
Firmographics- Name
- SafetyKit
- Legal name
- SafetyKit, Inc.
- Website
- https://safetykit.com
- Company type
- Private
- Founded year
- 2022
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- SafetyKit is an AI-native trust, safety, and compliance platform that replaces human risk reviewers with autonomous agents, serving Fortune 500 marketplaces, payment platforms, card networks, and AI platforms with multi-modal content moderation, fraud prevention, and merchant intelligence at scale.
- Ownership category
- akta.pro rank
SafetyKit industry classification
Industry- Product category
- Trust & Safety Software
- NAICS
- Investigation and Security Services (5616), Security Systems Services (56162)
- SIC
- Services-Prepackaged Software (7372)
- akta.pro primary industry
- Marketplace & Platform Trust & Safety (Buyer/Seller Abuse) (HDADALAI)
- akta.pro secondary industries
- Messaging, Chat & Live Interaction Safety (Text/Voice/Video) (BPAAALAC), Responsible AI, Security & Privacy Platforms (Safety, Guardrails, PII) (HDAEANAG), Fintech & Payments Content & Compliance Screening (KYC/AML-related content, fraud signals) (BPAAALAH), Risk Intelligence, Threat Monitoring & Crisis Response (Harmful trends, coordinated abuse) (BPAAALAN), UGC Moderation, Trust & Safety Platforms (MPACABAL)
Keywords
Where SafetyKit is headquartered
LocationHeadquarters
- HQ city
- San Francisco
- HQ country
- United States
- HQ region
- North America
Offices1 record
Markets served
SafetyKit business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Infrastructure, Marketing or Sales, Operations
Revenue model
- SaaS/API Platform Subscription: SafetyKit operates as a B2B SaaS platform providing AI-powered trust and safety infrastructure. Customers pay for access to the detection and decisioning engine, API usage, and policy enforcement capabilities. The platform processes customer data and returns enforcement decisions, labels, and risk scores through its API.
Go-to-market motion1 record
Distribution channels2 records
Marketing channels5 records
SafetyKit product offering
Product offeringCore offering
SafetyKit provides an AI-powered trust, safety, and compliance platform that replaces human risk and content reviewers with autonomous AI agents. The platform ingests text, images, video, audio, user events, transactions, and URLs in real time at 4.2k/s and uses LLMs to enforce 200+ pre-built policies covering content moderation, fraud prevention, merchant intelligence, and regulatory compliance (DSA, OSA, Take It Down Act) for enterprise platforms and marketplaces.
Product overview
SafetyKit is a unified AI-powered trust and safety platform consisting of a core detection and decisioning engine with modular add-on capabilities. The core platform ingests 100% of platform content (text, images, video, audio, user events, transactions, URLs) at 4.2k/s and provides built-in fraud ring network detection. The four primary solutions are Content Moderation with AI (for text, images, live video, and AI content), Fraud Prevention with AI (across transactions, merchants, and products), Merchant Intelligence (automated merchant investigations), and Networks (network-scale takedowns of fraud rings). Supporting the core engine, SafetyKit offers modular add-ons: Credit Risk (real-time credit loss prevention), Merchant Monitoring (continuous post-onboarding surveillance), AI Ops (autonomous AI agent workflows), ML Ops (custom classifiers, clustering, and anomaly detection), Human Ops (case management and reviewer tooling), Compliance Solutions (DSA, OSA, Take It Down Act coverage), and a Policy Library of 200+ pre-built policies. The entire platform is backed by a REST API and SDKs in TypeScript, Python, Kotlin, and Java.
Differentiator
Problem solved
Functional benefit
Products and services
- SafetyKit Platform (Core Detection & Decisioning Engine) The central detection and decisioning engine ingesting 100% of platform content and data (text, images, video, audio, user events, transactions, URLs) at 4.2k/s. Analyzes entire entities and their associated data to detect scams, fraud, and abuse, and surfaces enforcement insights and recommendations with <3200ms latency. Built-in network detection flags fraud rings, reused data elements, and repeated violations.
- Content Moderation with AI AI-powered moderation for text, images, live video, and AI content at scale using context-aware LLMs that apply written policies with precision. Enforces 200+ built-in policies covering hate speech, violence, CSAM, sexual content, harassment, misinformation, and more.
- Fraud Prevention with AI AI agents that detect and prevent fraud across transactions, merchants, and products at scale. Identifies payment fraud, synthetic identity fraud, triangulation fraud, fake reviews, phishing, refund abuse, promo code fraud, ticket fraud, marketplace scams, and AI-assisted fraud, integrating transaction, behavioral, and network-level signals.
- Merchant Intelligence Automated merchant investigations and intelligence at scale. AI agents investigate what merchants actually sell — across websites, social media, and storefronts — by following links, verifying documents, and analyzing product listings. Surfaces MCC mismatches, undisclosed products, and connections to known bad actors.
- Networks (Network-Scale Takedowns) Uncovers fraud rings hiding within platform data and removes them at scale. Detects coordinated fraud activity across accounts, devices, and behavioral patterns that individual-account analysis misses. Visualizes hidden relationships and maps fraud networks for enforcement action.
- Credit Risk Credit risk management solution that unifies transaction, content, and seller behavior signals to prevent credit losses before they occur. Provides dynamic ML-powered risk scoring, deep offsite investigations, real-time chargeback prevention, and fraud network mapping for enterprise credit teams, with reported 94% fraud detection rate and <0.1% false positive rate.
Quantifiable outcome
- Upwork achieves $1M+ annual savings vs manual review while maintaining 95%+ accuracy
- +5 more outcomes
Companies that use SafetyKit
Customer profileNamed customers11 records
Segments4 records
Ideal customer profiles3 records
SafetyKit technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration1 record
AI capability13 records
Feature8 records
SafetyKit partnerships and signals
Strategic signalPartnerships
Four partnerships are on record, tiered flagship and core.
- OpenAIflagshipSafetyKit partnered with OpenAI to support the release of gpt-oss-safeguard models (120b and 20b) designed to classify online safety harms. SafetyKit was an early testing partner during the research preview, helping validate the models for real-world moderation scenarios. The partnership demonstrates SafetyKit's role in advancing open AI safety infrastructure.
- DiscordflagshipDiscord was an early partner testing gpt-oss-safeguard models during the research preview, validating model flexibility in real-world moderation scenarios alongside SafetyKit.
- ROOSTcoreROOST established a Model Community to explore open AI models for online safety and collaborated with SafetyKit on early testing of gpt-oss-safeguard models.
- TomorocoreTomoro collaborated with OpenAI and SafetyKit on early testing of gpt-oss-safeguard models for online safety classification.
Scale indicators14 records
Recent moves6 records
Expansion highlights6 records
SafetyKit competitors and assessment
Company assessmentDirect peers
- ActiveFence: ActiveFence is the largest venture-backed trust & safety platform, offering content moderation, fraud detection, and threat intelligence for online platforms. It is the closest direct competitor to SafetyKit, sharing the same enterprise B2B SaaS model and Fortune 500 customer profile.
- Hive: Hive provides AI-powered content moderation APIs for text, image, video and audio, serving platforms and developers. It overlaps directly with SafetyKit's Global Ingestion Engine and pre-trained classifier offering.
- Spectrum Labs: Spectrum Labs (which absorbed Two Hat Security) provides AI-driven content moderation and trust & safety tools for online communities, gaming platforms and social networks. Comparable to SafetyKit's content moderation and policy enforcement products.
- Besedo: Besedo combines AI automation with human-in-the-loop moderation for marketplaces, dating apps and classifieds. Directly competes with SafetyKit for the freelance and marketplace use cases serving Upwork- and Faire-style customers.
- Cinder: Cinder offers a trust & safety operating system with policy generation, content moderation and case management tooling for large platforms. Closely mirrors SafetyKit's combined AI Ops + Human Ops + Policy Library stack.
- Forter: Forter is a digital commerce fraud prevention platform serving marketplaces and online retailers with real-time decisioning. Directly overlaps SafetyKit's Fraud Prevention and Marketplace Integrity use cases for buyer/seller abuse.
- Sift: Sift provides a digital trust and safety suite spanning fraud, content abuse and account protection for global platforms. Overlaps SafetyKit's fraud prevention, network detection and policy enforcement for enterprise customers.
- Arkose Labs: Arkose Labs provides bot management, account takeover prevention and fraud abuse protection for consumer-facing platforms. Directly comparable to SafetyKit's abuse prevention, network detection and marketplace integrity offerings.
Emerging players
- Unitary: Unitary applies multimodal AI to video content moderation at scale, with strong focus on livestream and user-generated video platforms. Overlaps with SafetyKit's video modality and is a meaningful emerging competitor in the same end market.
Broad incumbents
- Concentrix: Concentrix (which acquired the former ModSquad) operates a large-scale trust & safety BPO practice alongside its broader CX portfolio. It is the legacy incumbent SafetyKit is actively displacing for enterprise content moderation workloads.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat6 records
Key risks6 records
Key highlights7 records
Customer concentration
SafetyKit social profiles
Digital presenceSafetyKit financial estimates
Financial estimateRevenue estimate
Valuation estimate
SafetyKit leadership team
Management profileNumber of profiles
Profiles3 records
SafetyKit funding detail
Funding detailFunding overview
Funding rounds3 records
Investors4 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
SafetyKit M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about SafetyKit
What does SafetyKit do?
SafetyKit provides an AI-powered trust, safety, and compliance platform that replaces human risk and content reviewers with autonomous AI agents. The platform ingests text, images, video, audio, user events, transactions, and URLs in real time at 4.2k/s and uses LLMs to enforce 200+ pre-built policies covering content moderation, fraud prevention, merchant intelligence, and regulatory compliance (DSA, OSA, Take It Down Act) for enterprise platforms and marketplaces.
Is SafetyKit a public or private company?
SafetyKit is a private company. It is classified as venture growth investor backed and is currently operating.
When was SafetyKit founded?
SafetyKit was founded in 2022. It employs 11 to 50 people.
Where is SafetyKit based?
SafetyKit is headquartered in San Francisco, United States, in the North America region.
How does SafetyKit make money?
One revenue line is on record: saaS/API Platform Subscription.
Who are SafetyKit's main competitors?
Direct peers on record are ActiveFence, Hive, Spectrum Labs, Besedo, Cinder, Forter, Sift and Arkose Labs. Unitary is listed as an emerging player. Concentrix is listed as a broad incumbent.
Does SafetyKit have an API?
Yes. SafetyKit provides a REST API at https://api.safetykit.com/v1/ for ingesting and processing user, transaction, product, and merchant data across configurable namespaces. The API is public-facing and includes authentication via Bearer token (API key), asynchronous request/polling pattern, webhook delivery for real-time event notifications (workflow.succeeded, workflow.failed), upload URL flow for large batch payloads, and structured output including actions, labels, confidence scores, and label change diffs. Supported SDKs include TypeScript, Python, Kotlin, and Java. Webhook signature verification uses the Svix library with timestamp replay-attack protection. Developer documentation is at docs.safetykit.com.
What industry is SafetyKit in?
SafetyKit's product category is Trust & Safety Software. Its primary akta.pro industry code is HDADALAI, Marketplace & Platform Trust & Safety (Buyer/Seller Abuse), with a secondary code of BPAAALAC, Messaging, Chat & Live Interaction Safety (Text/Voice/Video). Its NAICS code is 5616 and its SIC code is 7372.