Agent Kernel

Agent Kernel is an open source platform for building and deploying enterprise AI agents at scale. It also describes itself as an operating system and an open-source runtime and orchestration layer for scalable, compliant enterprise AI agents.

The site highlights several concrete capabilities. Knowledge base support is built in for ChromaDB, Neo4j, and Starburst Galaxy, and a custom adapter API is provided to connect other backends. Agent Skills is presented as a set of CLI tools, including commands such as ak-init, ak-build, ak-add-capabilities, ak-add-integration, ak-cloud-deploy, and ak-test. The ak-init skill scaffolds a clean project structure and is described as working with any framework and any deployment target. The site also says Agent Skills works with coding tools such as Copilot, Claude, Cursor, and Windsurf. Security and compliance are mentioned as part of a certified ISO 27001 environment and an AICPA SOC 2 audit covering security, availability, and confidentiality.

Agent Kernel is described for several audiences: business leaders who want to incorporate AI agents into business workflows without handling technical complexity, developers who want to ship something robust without learning a new stack from scratch, and AI engineers who need a production-grade AI agent execution framework. It is delivered through documentation and a CLI installation path using pip, and the deployment section says the same agent code can run on AWS, Azure, GCP, or on-prem Docker without rewrites. Named deployment targets include AWS Lambda, AWS ECS/Fargate, Azure Functions, Azure Container Apps, and Google Cloud Run, with Cosmos DB session storage mentioned for Azure and Firestore session storage for Google Cloud. The site also mentions production-ready Terraform modules.

Pricing and licensing are stated as free, open source, and Apache 2.0, with no licensing costs and no vendor lock-in.