Create Collection
Learn how to add agent and MCP collections to Akto for centralized agentic security management.
You can add your agent and MCP collections to Akto to centralize and organize your agentic components. It allows you to organize your AI agents and MCP servers by function, making it easier to discover, monitor, and secure them. This helps maintain consistent security across your agentic systems and simplifies team collaboration. There are two methods to add your collections:
Create a New Collection Manually
Create Automated Collections via Traffic Connectors
Add a New Collection Manually
You can create a new empty collection and then populate it with agent components.
How to Create a New Collection
Navigate to Agentic Discovery > Collections
Click on the Create New Collection button
Enter a name for your new collection (e.g., "Production AI Agents", "MCP Servers - Customer Service")
Select the collection type (Agent Collection or MCP Collection)
Click on the Create button
After creating your collection, you can start monitoring agent traffic and running Agentic Security tests on your AI agents and MCP components.
Create Automated Collection
Akto automatically populates your agentic inventory by connecting to your traffic sources. This traffic source is called a traffic connector. Akto offers integration to multiple traffic connectors for capturing agentic system traffic.
Supported Traffic Connectors
AWS Traffic Mirroring
Connect Akto seamlessly with AWS Traffic Mirroring for comprehensive network visibility of your agentic systems. Gain real-time insights into AI agent traffic flow and enhance monitoring capabilities effortlessly. Click here for detailed steps.
AWS EKS
Easily integrate Akto with AWS EKS for streamlined container orchestration. Manage and scale your containerized AI agents effortlessly within the AWS Kubernetes environment. For detailed steps, click here.
AWS Fargate
Effortlessly link Akto with AWS Fargate for seamless serverless container management. Streamline deployment and scaling of agent containers without managing the underlying infrastructure. For detailed steps, click here.
AWS Beanstalk
Connect Akto seamlessly with AWS Elastic Beanstalk for simplified agentic application deployment and management. Easily scale and monitor your AI agents within the AWS environment. For more information, please refer to the Connect Akto with AWS Beanstalk documentation.
Kubernetes in AWS
You can add Akto DaemonSet to your Kubernetes cluster. It is very lightweight and will run 1 per node in your Kubernetes cluster. It will intercept all the node traffic (including agentic traffic) and send it to the Akto traffic analyzer. For detailed information, please refer to the Connect Akto with Kubernetes in AWS documentation.
AWS API Gateway
You can connect Akto with AWS API Gateway for automated agent component detection, as it simplifies the process of discovering, accessing, and integrating a wide range of agent endpoints and MCP servers. Please refer to the Connect Akto with AWS API Gateway documentation.
Kong
You can connect Akto with Kong for automated agent detection which enhances the platform's agility, accelerates integration, improves management, and ensures a more secure and efficient ecosystem for developers leveraging agentic systems. Click here for detailed steps.
NGINX
You can effortlessly connect Akto with NGINX for enhanced web server performance and reliability. Optimize traffic handling for your agentic systems and boost your application's speed with this seamless integration. Please refer to the Connect Akto with NGINX documentation for detailed steps.
Envoy
Integrate Akto seamlessly with Envoy for advanced service mesh capabilities. Enhance traffic management and secure communication between agentic services effortlessly. Check out the Connect Akto with Envoy documentation for detailed information.
Docker
You can effortlessly link Akto with Docker for streamlined containerization by adding the Akto-traffic-collector service to your docker-compose file. Perfect for monitoring containerized AI agents. Please refer to the Connect Akto with Docker documentation for detailed steps.
TCP Agent
Connect Akto seamlessly with TCP Agent for efficient TCP traffic monitoring and analysis. Enhance visibility and control over agentic network communication effortlessly. Click here for detailed steps.
eBPF
You can connect Akto with eBPF for automated agent detection as it offers an immediate understanding of how AI agents are used, ensures a complete list of all available agent components, boosts security, and helps Akto adjust quickly to changes in your agentic ecosystem. This connection makes Akto work better and more smoothly overall. Click here for the detailed steps.
GCP Packet Mirroring
Packet Mirroring is a GCP feature that clones the traffic of specified instances in your Virtual Private Cloud (VPC) network and forwards it to Akto. Packet Mirroring captures all agentic traffic and packet data, including payloads and headers. The capture can be configured for both egress and ingress traffic. For detailed information, click here.
Best Practices
Organize collections by environment (production, staging, development)
Separate agent collections from MCP collections for clarity
Use descriptive names that reflect the collection's purpose
Regularly review and update collection memberships
Monitor collection growth to identify new agent components
Last updated