Enkrypt AI

AI security platform for agent red teaming, runtime guardrails, policy enforcement, and MCP risk assessment, with free evaluation access and paid credit-based plans.

About Enkrypt AI

Enkrypt AI provides security testing and runtime controls for AI applications and agents. Its product range includes Agent Red Teaming, Agent Guardrails, Agent Policy Engine, and MCP Scanner. Together, these capabilities help teams identify weaknesses, apply controls to live interactions, and maintain evidence of security decisions.

Red teaming evaluates systems that combine models, retrieval, tools, and agent workflows. Guardrails can approve, modify, or block risky behavior at several points in that workflow, including inputs, retrieved content, tool calls, and outputs. The policy engine translates organizational requirements into controls linked to their source requirements.

MCP Scanner examines server capabilities and permissions, then produces findings and remediation guidance. This adds a component-level assessment to the platform’s behavioral testing and runtime enforcement.

Key Features

Agent red teaming: Tests applications, tools, retrieval workflows, and MCP interactions against adversarial scenarios.
Runtime guardrails: Applies controls at prompt, retrieval, tool, and output boundaries.
Tool-action enforcement: Supports approval or denial of actions, tool allowlists, and restrictions on connected resources.
Policy-to-control mapping: Converts policy text and documents into controls with source traceability and review workflows.
MCP risk assessment: Inventories server capabilities and flags excessive permissions, sensitive access, and unsafe actions.
Audit evidence: Records enforcement decisions and supports exports for security and governance reviews.

Pricing

Explore: $0/month, with 500 starting credits and 50 credits/month. No credit card required.
Launch: $149/month billed monthly, or $134/month billed annually. Includes 5,000 starting credits and 250 credits/month.
Scale: $1,499/month billed monthly, or $1,349/month billed annually. Includes 10,000 starting credits and 1,000 credits/month.
Enterprise: Custom pricing; contact sales. Custom credits, VPC/on-premises deployment, and unlimited VPC usage are listed.
Additional usage: Extra credits are available at the applicable tier rate; numeric rates are not published. A comprehensive red-team assessment requires 5,000 credits.
Add-ons: Dedicated customer-success and enhanced support options appear as add-ons; prices are unspecified.
Plan changes: Upgrades are prorated; downgrades apply at the next billing cycle. Conflicting retention, rollover, and feature-table entries are excluded.

Use Cases

Test AI applications for failures involving prompts, retrieval, and connected tools.
Enforce security rules before agent actions execute.
Translate internal AI policies into controls with traceable evidence.
Assess MCP servers before granting agents access to their capabilities.

Pros & Cons

Pros:

Connects adversarial findings with guardrail policies and operational monitoring.
Applies controls across several stages of an agent workflow.
Supports policy traceability and exportable evidence for reviews.

Cons: