Akamai Firewall for AI claims input and output controls that prevent sensitive-data leakage in AI applications.
Applies multilayered input and output guardrails to prevent sensitive data exposure
Public-source research
Vendor evidence
Review what vendors say publicly, the exact quoted source text, the security requirement each statement may support, and what still needs verification.
Research library coverageCounts describe the research workflow, not vendor quality or product effectiveness.View details
Evidence records1280
Source checked863
Needs verification0
Source captured0
No supporting claim found411
Excluded from evidence6
Showing 1–20 of 70 evidence records
Clear all filtersSalt claims sensitive-data flow visibility and blocking of data exfiltration through agent-connected application programming interfaces (APIs).
Block data exfiltration & attacks
Imperva claims detection, blocking, masking, and alerting for sensitive data in model inputs and outputs.
block, mask, or alert on accidental exposure of PII and sensitive data
Kong claims built-in PII sanitization that prevents sensitive data from reaching an large language model (LLM).
sensitive data never leaks into an LLM
Speakeasy claims real-time blocking, redaction, or logging of sensitive data in prompts, responses, and agent actions.
Sensitive data is blocked, redacted, or logged
Harness claims inspection of AI prompts and responses for PII and other sensitive data.
inspect AI prompts and responses for sensitive data types
Wallarm claims agentic AI payload protection against injection attacks and data leakage.
preventing injection attacks and data leakage
Cequence claims data loss prevention (DLP) scanning, monitoring, redaction, and blocking across AI-agent requests and Model Context Protocol (MCP) responses.
Apply DLP scanning to AI agent requests and MCP server responses
Upwind claims real-time detection of sensitive data in prompts and inference payloads traversing cloud networks.
Real-time detection of sensitive data in prompts and inference payloads
Proofpoint claims real-time inspection of prompts, outputs, and agent workflows to identify and block sensitive-data exposure.
Enterprises can reduce sensitive data exposure by inspecting prompts, outputs, and agent workflows in real time.
Securiti claims mapping and monitoring of data sources, processing, risks, obligations, and data flows connected to AI models.
Map AI models to data sources, monitor data flow
Snyk Evo materials reviewed did not provide a public claim for inspection or enforcement over sensitive data moving through AI.
No quoted source text is recorded for this claim.
ModelOp materials reviewed did not provide a public claim for inspection or enforcement over sensitive data moving through AI.
No quoted source text is recorded for this claim.
LatticeFlow AI materials reviewed did not provide a public claim for inspection or enforcement over sensitive data moving through AI.
No quoted source text is recorded for this claim.
Runlayer claims pre-action runtime scanning of tool calls, outputs, intent, and sensitive data.
scan tool calls, outputs, intent, and sensitive data before risky actions reach company systems
Keycard materials reviewed did not provide a public claim for inspection or enforcement over sensitive data moving through AI.
No quoted source text is recorded for this claim.
AWS claims Bedrock does not store or use customer data to train models and supports encryption and identity-based access policies.
Bedrock never stores or uses your data to train models, ensuring complete security and privacy, with encryption of data in transit and at rest, as well as identity-based policies for managing data access.
Google claims Model Armor can mitigate leakage of sensitive IP and PII in large language model (LLM) prompts or responses.
Mitigate the risk of leaking sensitive intellectual property (IP) and personally identifiable information (PII) in LLM prompts or responses.
Microsoft claims Foundry guardrail controls cover protected material and personally identifiable information for models and agents.
Protected material for code ✅ ✅ Protected material for text ✅ ✅ Groundedness (Preview) ✅ ❌ Personally identifiable information (Preview) ✅ ✅
Zscaler claims AI Access Security blocks sensitive data loss in prompts using inline data loss prevention (DLP) dictionaries for source code, PII, PCI, PHI, and other data types.
Block the loss of sensitive data in prompts with powerful inline DLP across 100+ DLP dictionaries like Source Code, PII, PCI, PHI, and more.