Gray Swan

Gray Swan sells security for AI applications: Shade, which runs autonomous adversarial campaigns against AI agents, Cygnal, which filters model inputs and outputs, and the Arena red-teaming network for model builders.

StageEarly-stage
Founded—
Headquarters—
StatusPrivate
grayswan.ai

Products

Selected offerings — the AI products recorded here, not the vendor's full catalogue.

Similar products 150

No sourced competitor yet: no filing, vendor document or measured-market report read so far names a rival for this company, so none is asserted. The companies below are the closest by what their products do.

How this list is ranked ↗

GiskardUndisclosedCygnal and Giskard Guards both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareRepello AIUndisclosedShade and ARTEMIS both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareNomaUndisclosedCygnal and Noma Agent Access Control both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareFabraixUndisclosedShade and Nyx both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareAsenionUndisclosedShade and Asenion Verify both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; scale not recorded for either.2 of 2 capabilitiesShared product typeComparepromptfooAcquiredShade and Promptfoo both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; Gray Swan's scale not recorded; promptfoo is acquired, with no independent scale.2 of 2 capabilitiesShared product typeCompareAdvaiUndisclosedShade and Advai Platform both list guardrails and safety; both also serve buyers who need to use AI to attack our own systems first; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareMindgardEarly-stageShade and Mindgard AI Security Platform both list guardrails and safety; both also serve buyers who need to use AI to attack our own systems first; same layer and the same scale band (early-stage private).2 of 2 capabilitiesShared product typeSame scale bandCompareHamming AIEarly-stageShade and Hamming both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareQualifireUndisclosedShade and Qualifire both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareLassoEarly-stageShade and Lasso AI Security Platform both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; same layer and the same scale band (early-stage private).2 of 2 capabilitiesShared product typeSame scale bandCompareAdversa AIUndisclosedShade and AI Red Teaming both list guardrails and safety; both also serve buyers who need to use AI to attack our own systems first; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareCraniumEarly-stageShade and Cranium ONE both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; same layer and the same scale band (early-stage private).2 of 2 capabilitiesShared product typeSame scale bandCompareGlow SecurityUndisclosedCygnal and Glow both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareLumia SecurityEarly-stageCygnal and AI Usage Control both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareOnyx SecurityUndisclosedCygnal and Onyx both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareZscalerPublicShade and Zscaler AI Red Teaming both list guardrails and safety; both also serve buyers who need to use AI to attack our own systems first; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeComparePalo Alto NetworksPublicShade and Prisma AIRS AI Red Teaming both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareVijilEarly-stageCygnal and Vijil Dome both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; ships AI agent, application and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerSame scale bandCompareSingulr AIEarly-stageCygnal and Singulr Runtime Governance both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareLiminalUndisclosedCygnal and Liminal both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareHarmonic SecurityEarly-stageCygnal and Harmonic Guide both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; ships application rather than the same layer.1 of 2 capabilitiesDifferent layerSame scale bandCompareLogosGuardUndisclosedCygnal and LogosGuard Enterprise Browser Extension both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either; ships application and infrastructure service rather than the same layer.1 of 2 capabilitiesDifferent layerCompareMicrosoftPublicShade and AI Red Teaming Agent both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareInvisible TechnologiesLate-stageShade and Synapse both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareTraceforceUndisclosedCygnal and Traceforce both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareNeuralTrustEarly-stageCygnal and TrustGuard both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either; ships developer tool and infrastructure service rather than the same layer.2 of 2 capabilitiesDifferent layerCompareSAPPublicShade and SAP AI Agent Hub both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareMetorialUndisclosedCygnal and Protoguard both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareDynamo AIUndisclosedShade and AgentWarden both list guardrails and safety; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareSpeakeasyUndisclosedCygnal and Agent Identity both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareEnkrypt AIAcquiredShade and Enkrypt AI both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; Gray Swan's scale not recorded; Enkrypt AI is acquired, with no independent scale; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompareDataikuLate-stageShade and Dataiku Agent Management both list guardrails and safety; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareIndora LabsUndisclosedCygnal and Arbiter both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either; ships application rather than the same layer.1 of 2 capabilitiesDifferent layerCompareCyeraLate-stageCygnal and Cyera Agent Guardian both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded.1 of 2 capabilitiesShared product typeCompareFuture AGIUndisclosedShade and Future AGI both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareMaxim AIUndisclosedCygnal and Bifrost Edge both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareObotUndisclosedShade and Obot AI Control Plane both list guardrails and safety; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareGlean (Glean Technologies)Late-stageCygnal and Glean AI Gateway both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded.1 of 2 capabilitiesShared product typeCompareIBMPublicShade and IBM DataPower Interact Gateway both list guardrails and safety; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeComparePillar SecurityEarly-stageCygnal and Agentic Endpoint both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareWorkdayPublicCygnal and Agent Passport both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareHiddenLayerLate-stageCygnal and HiddenLayer AI Runtime Security both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareWriterLate-stageCygnal and WRITER AI Studio both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded.1 of 2 capabilitiesShared product typeCompareDreadnodeEarly-stageShade and Dreadnode Platform both serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareLumenova AIUndisclosedCygnal and AI Guardrails both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either; ships infrastructure service rather than the same layer.2 of 2 capabilitiesDifferent layerCompareStackOneEarly-stageCygnal and StackOne Defender both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; ships application and developer tool rather than the same layer.1 of 2 capabilitiesDifferent layerSame scale bandCompareVerifyWiseUndisclosedCygnal and Shadow AI Detection both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareArcadeUndisclosedCygnal and Arcade both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareGeordie AIEarly-stageCygnal and Beam both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; ships application rather than the same layer.2 of 2 capabilitiesDifferent layerSame scale bandCompareCheck Point Software TechnologiesPublicShade and AI Red Teaming both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareJFrogPublicCygnal and JFrog Agent Guard both list guardrails and safety; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareMergeEarly-stageCygnal and Merge for Workforce both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; same layer and the same scale band (early-stage private).1 of 2 capabilitiesShared product typeSame scale bandCompareArthur AIEarly-stageShade and Arthur Agent Development Toolkit both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent; same layer and the same scale band (early-stage private).2 of 2 capabilitiesShared product typeSame scale bandCompareAembitUndisclosedCygnal and IAM for Agentic AI both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareFiddler AILate-stageCygnal and Fiddler Guardrails both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareMintiveEarly-stageShade and Mintive both list guardrails and safety; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareAurascapeEarly-stageCygnal and Aurascape AI Security Platform both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; same layer and the same scale band (early-stage private).1 of 2 capabilitiesShared product typeSame scale bandCompareRunlayerEarly-stageCygnal and Runlayer MCP Gateway both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; same layer and the same scale band (early-stage private).1 of 2 capabilitiesShared product typeSame scale bandCompareObiguardUndisclosedShade and Governance AI both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareWitnessAIEarly-stageCygnal and Witness Protect both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; same layer and the same scale band (early-stage private).1 of 2 capabilitiesShared product typeSame scale bandCompareZuploUndisclosedCygnal and MCP Gateway both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareLatticeFlow AIUndisclosedShade and LatticeFlow AI Evaluate both serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareSentinelOnePublicCygnal and Prompt Security both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded.1 of 2 capabilitiesShared product typeCompareKnosticEarly-stageCygnal and Kirin both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either; ships application, data service and 1 more rather than the same layer.1 of 2 capabilitiesDifferent layerCompareAuthMindUndisclosedCygnal and Agentic AI Identity Protection both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareCoArenaUndisclosedShade and Introspect both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent; scale not recorded for either.2 of 2 capabilitiesShared product typeComparePreambleUndisclosedCygnal and AI Threat Detection both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareDescopeEarly-stageCygnal and Agentic Identity Hub both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareZenityLate-stageCygnal and Zenity Platform both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeComparePortUndisclosedCygnal and Governance both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeComparePayman AIEarly-stageShade and Payman governed agent layer both list guardrails and safety; scale not recorded for either; ships AI agent rather than the same layer.1 of 2 capabilitiesDifferent layerComparePermit.ioUndisclosedShade and Permit MCP Gateway both list guardrails and safety; scale not recorded for either; ships developer tool rather than the same layer.2 of 2 capabilitiesDifferent layerCompareSalesforcePublicCygnal and MuleSoft Agent Fabric both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded; ships AI agent, API service and 3 more rather than the same layer.1 of 2 capabilitiesDifferent layerCompareCredalUndisclosedShade and Agent Registry both list guardrails and safety; scale not recorded for either; ships agent platform rather than the same layer.2 of 2 capabilitiesDifferent layerCompareServiceNowPublicShade and AI Control Tower both list guardrails and safety; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeComparePrediction GuardEarly-stageCygnal and Prediction Guard Control Plane both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareTruthScanUndisclosedCygnal and Image Content Moderation both list guardrails and safety; scale not recorded for either; ships API service and application rather than the same layer.1 of 2 capabilitiesDifferent layerCompareValidaitorUndisclosedShade and Trustworthy AI Platform both serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareHippocratic AILate-stageShade and AI Call Supervisor both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent; Gray Swan's scale not recorded; ships AI agent, agent platform and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompareCredo AIEarly-stageCygnal and Agent Governor both list guardrails and safety; ships AI agent and application rather than the same layer.1 of 2 capabilitiesDifferent layerSame scale bandCompareUpsonicUndisclosedShade and AgentOS both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent; scale not recorded for either; ships AI agent and infrastructure service rather than the same layer.2 of 2 capabilitiesDifferent layerCompareAnacondaLate-stageShade and AI Security & Guardrails both list guardrails and safety; both also serve buyers who need to use AI to attack our own systems first; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareDatadogPublicCygnal and Datadog AI Guard both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded; ships AI agent, API service and 3 more rather than the same layer.2 of 2 capabilitiesDifferent layerComparexpander.aiEarly-stageCygnal and AI Gateway both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareNVIDIAPublicShade and NVIDIA Open Agent Safety Platform both list guardrails and safety; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareAlethicaUndisclosedShade and Policy-as-Code both list guardrails and safety; scale not recorded for either; ships AI agent and agent platform rather than the same layer.1 of 2 capabilitiesDifferent layerCompareValidMindUndisclosedCygnal and ValidMind Agent Authority both list guardrails and safety; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareAlinia AIEarly-stageShade and Seny both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent; scale not recorded for either; ships model API rather than the same layer.2 of 2 capabilitiesDifferent layerCompareAgentic FabriqUndisclosedCygnal and Agentic Fabriq both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either; ships agent platform and developer tool rather than the same layer.1 of 2 capabilitiesDifferent layerCompareDecaworkUndisclosedCygnal and Decawork both list guardrails and safety; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareAutomation AnywhereUndisclosedShade and AI Governance both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareJasperUndisclosedCygnal and Jasper Brand IQ both list guardrails and safety; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareRogoLate-stageShade and Watchtower both list guardrails and safety; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareEnzaiUndisclosedCygnal and AI discovery and inventory both serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareParloaLate-stageShade and Parloa Lens both list guardrails and safety; Gray Swan's scale not recorded; ships AI agent, agent platform and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompareKeycardUndisclosedCygnal and Keycard for Code both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either; ships infrastructure service rather than the same layer.1 of 2 capabilitiesDifferent layerCompareStacktraceUndisclosedCygnal and Stacktrace Personal Agent Gateway both list guardrails and safety; scale not recorded for either; ships agent platform and developer tool rather than the same layer.2 of 2 capabilitiesDifferent layerCompareCodeRabbitLate-stageCygnal and CodeRabbit Security both list guardrails and safety; Gray Swan's scale not recorded; ships AI agent and developer tool rather than the same layer.1 of 2 capabilitiesDifferent layerCompareAmazonPublicCygnal and Amazon Bedrock Guardrails both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareWizAcquiredCygnal and Wiz AI-SPM both serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded; Wiz is acquired, with no independent scale.Shared product typeCompareKlira AIUndisclosedShade and Klira SDK both list guardrails and safety; scale not recorded for either; ships developer tool rather than the same layer.2 of 2 capabilitiesDifferent layerCompareLatentUndisclosedShade and Latent both list guardrails and safety; scale not recorded for either; ships developer tool rather than the same layer.2 of 2 capabilitiesDifferent layerComparenRouterUndisclosedShade and nRouter LLM Gateway both list guardrails and safety; scale not recorded for either; ships API service rather than the same layer.2 of 2 capabilitiesDifferent layerCompareTrueFoundryEarly-stageShade and TrueFoundry MCP Gateway both list guardrails and safety; same layer and the same scale band (early-stage private).2 of 2 capabilitiesShared product typeSame scale bandCompareSASUndisclosedCygnal and SAS AI Navigator both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded; SAS's scale not comparable: only a revenue estimate, no market cap or valuation.2 of 2 capabilitiesShared product typeCompareStraikerEarly-stageShade and Ascend AI both serve buyers who need to test and evaluate an AI agent and use AI to attack our own systems first; scale not recorded for either; ships application rather than the same layer.2 of 2 capabilitiesDifferent layerCompareSkyfireUndisclosedCygnal and Know Your Agent (KYA) both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either; ships infrastructure service rather than the same layer.1 of 2 capabilitiesDifferent layerCompareAshrUndisclosedShade and Ashr Evals both serve buyers who need to test and evaluate an AI agent; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareHarnessLate-stageCygnal and AI Posture both list guardrails and safety; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareNuPlay AIEarly-stageShade and NuLoop both serve buyers who need to test and evaluate an AI agent; same layer and the same scale band (early-stage private).1 of 2 capabilitiesShared product typeSame scale bandCompareCUBIGUndisclosedCygnal and LLM Capsule both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareNeysaUndisclosedCygnal and Neysa Aegis LLM Shield both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareComposioUndisclosedCygnal and Firewall both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareSightengineUndisclosedCygnal and Sightengine Audio Moderation both list guardrails and safety; scale not recorded for either; ships API service rather than the same layer.1 of 2 capabilitiesDifferent layerCompareZapierLate-stageCygnal and AI Guardrails by Zapier both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded.1 of 2 capabilitiesShared product typeCompareDatabricksLate-stageCygnal and Unity Gateway both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareAll Hands AIEarly-stageShade and Agent Control Plane both list guardrails and safety; same layer and the same scale band (early-stage private).2 of 2 capabilitiesShared product typeSame scale bandComparePointFiveEarly-stageCygnal and TokenShift both serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareAnchorheadUndisclosedShade and Anchorhead both serve buyers who need to test and evaluate an AI agent; scale not recorded for either; ships application rather than the same layer.1 of 2 capabilitiesDifferent layerCompareTesslLate-stageShade and Tessl Registry both list guardrails and safety; scale not recorded for either; ships AI agent, application and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompareOrq.aiUndisclosedShade and AI Gateway both list guardrails and safety; scale not recorded for either; ships API service rather than the same layer.2 of 2 capabilitiesDifferent layerCompareLiteLLMEarly-stageShade and AI Gateway (LiteLLM Proxy) both list guardrails and safety; scale not recorded for either; ships AI agent, application and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerCompareScalekitUndisclosedCygnal and Agent Gateway both list guardrails and safety; both also serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareSierraLate-stageShade and Insights both serve buyers who need to test and evaluate an AI agent; Gray Swan's scale not recorded; ships AI agent, application and 1 more rather than the same layer.1 of 2 capabilitiesDifferent layerCompareSalusUndisclosedCygnal and Salus both list guardrails and safety; scale not recorded for either; ships developer tool rather than the same layer.1 of 2 capabilitiesDifferent layerCompareRobocurveUndisclosedShade and Robocurve evaluations both serve buyers who need to test and evaluate an AI agent; scale not recorded for either; ships data service rather than the same layer.1 of 2 capabilitiesDifferent layerCompareShaipUndisclosedShade and Shaip AI data services both serve buyers who need to test and evaluate an AI agent; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareSplunkAcquiredShade and Agent and Model Evaluation both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent; Gray Swan's scale not recorded; Splunk is acquired, with no independent scale.2 of 2 capabilitiesShared product typeCompareResemble AIEarly-stageCygnal and Resemble Watermarker both list guardrails and safety; ships API service and application rather than the same layer.1 of 2 capabilitiesDifferent layerSame scale bandCompareRequestyEarly-stageShade and Requesty AI Gateway both list guardrails and safety; scale not recorded for either; ships API service rather than the same layer.2 of 2 capabilitiesDifferent layerCompareOpenHackUndisclosedShade and AI Pentesting both serve buyers who need to use AI to attack our own systems first; scale not recorded for either; ships AI agent rather than the same layer.Different layerCompareVectaraEarly-stageShade and Vectara Guardian Agents both list guardrails and safety; ships AI agent, agent platform and 2 more rather than the same layer.2 of 2 capabilitiesDifferent layerSame scale bandCompareGuild.aiUndisclosedShade and Guild agent control plane both list guardrails and safety; scale not recorded for either; ships agent platform and developer tool rather than the same layer.2 of 2 capabilitiesDifferent layerCompareTolokaUndisclosedShade and Toloka Arena both serve buyers who need to test and evaluate an AI agent; scale not recorded for either.2 of 2 capabilitiesShared product typeCompareBroadcomPublicShade and AgentMinder both list guardrails and safety; Gray Swan's scale not recorded.1 of 2 capabilitiesShared product typeCompareKore.aiLate-stageShade and Arch AI both serve buyers who need to test and evaluate an AI agent; scale not recorded for either; ships agent platform and application rather than the same layer.1 of 2 capabilitiesDifferent layerCompareCiscoPublicShade and Cisco AI Defense both list guardrails and safety; both also serve buyers who need to test and evaluate an AI agent; Gray Swan's scale not recorded.2 of 2 capabilitiesShared product typeCompareDecagonLate-stageShade and Experiments both serve buyers who need to test and evaluate an AI agent; Gray Swan's scale not recorded; ships AI agent, agent platform and 1 more rather than the same layer.2 of 2 capabilitiesDifferent layerComparePalantirPublicShade and Palantir AIP both serve buyers who need to test and evaluate an AI agent; Gray Swan's scale not recorded.1 of 2 capabilitiesShared product typeCompareVercelLate-stageCygnal and Vercel Passport both list guardrails and safety; Gray Swan's scale not recorded; ships AI agent, API service and 2 more rather than the same layer.2 of 2 capabilitiesDifferent layerComparePromptFluentUndisclosedCygnal and PromptFluent both serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.Shared product typeCompareOktaPublicCygnal and Okta for AI Agents both list guardrails and safety; Gray Swan's scale not recorded; Okta's scale not comparable: only a revenue estimate, no market cap or valuation.1 of 2 capabilitiesShared product typeCompareConfident AIUndisclosedShade and AI Red Teaming both serve buyers who need to use AI to attack our own systems first; scale not recorded for either.1 of 2 capabilitiesShared product typeCompareHatz AIUndisclosedCygnal and Hatz AI both serve buyers who need to block unsafe AI inputs and unapproved AI use; scale not recorded for either.Shared product typeCompareHueUndisclosedShade and Hue both serve buyers who need to test and evaluate an AI agent; scale not recorded for either; ships developer tool rather than the same layer.1 of 2 capabilitiesDifferent layerCompareLemmaUndisclosedShade and Lemma both serve buyers who need to test and evaluate an AI agent; scale not recorded for either; ships developer tool rather than the same layer.1 of 2 capabilitiesDifferent layerCompareLivePersonAcquiredShade and LivePerson Syntrix both serve buyers who need to test and evaluate an AI agent; Gray Swan's scale not recorded; LivePerson is acquired, with no independent scale; ships AI agent, agent platform and 2 more rather than the same layer.1 of 2 capabilitiesDifferent layerCompareTolmoUndisclosedShade and Tolmo Pentesting Agent both serve buyers who need to use AI to attack our own systems first; scale not recorded for either; ships AI agent rather than the same layer.Different layerCompareByteDanceUndisclosedShade and Coze Loop both serve buyers who need to test and evaluate an AI agent; Gray Swan's scale not recorded; ByteDance's scale not comparable: only a revenue estimate, no market cap or valuation.1 of 2 capabilitiesShared product typeCompare

Closest alternatives are pairs a vendor page itself compares. Similar products are matched from what each product does; nothing in that group is asserted as a rivalry.

Sources

Stage, from total funding

News reporting · 28 May 2026 at least

Dealroom news, 2026-05-28: Gray Swan "has raised a $40M Series A", with Magarac Venture Partners named as an existing investor. A floor: only the Series A is sourced and earlier backing exists. Read through a page-reading tool; the site refuses scripted fetches.

Valuation

Not found · 11 Nov 2026

Pages read for this record state no valuation or market capitalisation.

Founded

Not found · 11 Nov 2026

Pages read for this record state no founding year.

Public listing

Not found · 11 Nov 2026

No listing on the pages read; treated as private.

Description

Official documentation · 11 Nov 2026

Company page: "Block adversarial inputs and unsafe outputs in real time"

Open Gray Swan in the directory

Something wrong here? Send a correction — quote this company id: gray-swan.