Guardrails for Every AI Coding Workflows
Akto now intercepts prompts, tool calls, and responses across Codex CLI and Neovim IDE. Enforce guardrails in real time, right where your developers build.
AI Discovery
Skills Inventory and Blocking
Discover all agent skills installed on your endpoints. Block risky skills in bulk from Skill Collections or enforce per-endpoint control.
AI Discovery
Endpoint-Level Insights
Map agentic assets to individual users and devices. Track risk scores, sensitive data exposure, and access patterns per endpoint.
AI Discovery
Discover Locally Hosted Models
Discover and secure on-prem models running on SGLang, vLLM, Docker Model Runner, and Ollama.
AI Discovery
Host Exclusion Controls
Exclude specific hosts from Traffic Analysis in Settings. Prevent internal or non-critical traffic from being captured and analyzed.
AI Discovery
SentinelOne Connector
Discover and monitor SentinelOne managed endpoints from your environment. Enforce real-time guardrails.
AI Discovery
Anthropic Connector
Capture traffic and surface real-time insights from Anthropic models directly in your Atlas dashboard.
AI Governance
MCP Registry Control
Import MCP registries via CSV URL to allowlist approved servers. Only listed servers are permitted for controlled and validated MCP access.
AI Governance
MCP Registry Sync
Add discovered MCP servers to registries directly from Audit Data, where they are automatically marked for easy identification.
AI Governance
Block personal accounts
Enable policies to enforce corporate-only access, blocking all personal and non-corporate accounts.
AI Guardrails
Response Guardrails
Validate and enforce security policies on response payloads across both direct and proxied requests at every endpoint.
AI Guardrails
File and URL Guardrails
Validate uploaded files and external URLs including audio, video, images, and documents.
AI Guardrails
PII Count-Based Guardrails
Monitor and control sensitive data exposure based on occurrence thresholds. Flag when PII counts exceed policy limits.
AI Guardrails
Alert-Only Guardrails
Configure guardrail policies that alert and warn on suspicious activity without blocking requests, for phased rollouts.
AI Guardrails
Apply Guardrails to All Servers
Apply guardrail policies across all agentic assets in a single action with the new "Apply to all" option.
Integrations
Codex CLI Hooks
Enforce Akto guardrails on prompts, tool calls, and responses in Codex-powered code generation workflows directly from the CLI.
Integrations
Neovim Hooks
Enforce real-time guardrails directly in Neovim with Akto hooks for prompt and response validation as you code.
Platform
HTTP/HTTPS Proxy Support
Configure HTTP/HTTPS proxies in Settings to deploy Akto in restricted network environments.
Platform
Custom Roles with Okta SSO
Create fine-grained roles (Admin, Developer, Threat Engineer, Threat Viewer) tailored to your team structure with Okta SSO.
Platform
GitHub Copilot for Ask Akto
Run Ask Akto with GitHub Copilot as the underlying AI model.
AI Guardrails
Response Guardrails
Validate and enforce security policies on response payloads across both direct and proxied requests at every endpoint.
AI Guardrails
File and URL Guardrails
Validate uploaded files and external URLs including audio, video, images, and documents.
AI Guardrails
PII Count-Based Guardrails
Monitor and control sensitive data exposure based on occurrence thresholds. Flag when PII counts exceed policy limits.
AI Guardrails
Alert-Only Guardrails
Configure guardrail policies that alert and warn on suspicious activity without blocking requests, for phased rollouts.
AI Guardrails
Apply Guardrails to All Servers
Apply guardrail policies across all agentic assets in a single action with the new "Apply to all" option.
AI Red Teaming
Tool Invocation Visibility
View which tools agents invoked during probe execution to trace scan behavior and identify exposure points.
AI Red Teaming
Category-wise scan scoring
View scan coverage by agentic risk category. Spot gaps and track pass/fail rates at a glance.
Reporting
Extended Compliance Mapping
Track compliance posture across EU AI Act, OWASP Agentic Top 10, NIST AI RMF, and Mitre ATLAS from a unified Compliance Page.
Integrations
Claude Agent SDK
Enforce guardrails on prompts, tool calls, and responses in real time via Claude Agent SDK's native callback system.
Platform
Expired Token Alerts
Spot expired module authentication tokens instantly with a visual red alert icon and tooltip in Settings.
Platform
HTTP/HTTPS Proxy Support
Configure HTTP/HTTPS proxies in Settings to deploy Akto in restricted network environments.
Platform
Custom Roles with Okta SSO
Create fine-grained roles (Admin, Developer, Threat Engineer, Threat Viewer) tailored to your team structure with Okta SSO.
Platform
GitHub Copilot for Ask Akto
Run Ask Akto with GitHub Copilot as the underlying AI model.