Wanaku, from Wanaku AI, is an MCP server tool and governed action proxy for AI agents. Built using Java, Apache Camel, Quarkus, and Rust, it connects AI models to enterprise systems. It provides a standardized gateway for AI interactions, supporting multi-agent orchestration and dynamic tool registration. Available via shell script, container images, or JBang, it aims to standardize AI access for developers building AI applications.

What does Wanaku connect to?

Wanaku serves as an MCP server, linking AI agents, including Large Language Models, to diverse enterprise systems. It integrates with cloud services, REST APIs, and databases. Furthermore, it can connect to other MCP servers, acting as a central hub. This connectivity allows AI agents to interact with existing infrastructure safely and effectively, standardizing access patterns for various AI clients like Claude Desktop or LangFlow.

Core features and tools

Wanaku enables integration developers to expose Apache Camel routes as callable tools for AI agents. It handles core MCP methods such as tools/list, tools/call, resources/list, resources/read, prompts/list, and prompts/get. Dynamic tool registration is supported by dropping YAML schemas into a directory and calling "/v1/tools/reload". It supports HTTP, WebSocket at "/v1/stream", and SSE protocols for real-time bidirectional communication.

Who is it for?

Wanaku targets integration developers and those building AI applications. It helps enterprises standardize AI access across teams. Concrete workflows include exposing Apache Camel routes as AI tools and facilitating multi-agent orchestration. It also functions as an enterprise AI gateway, unifying access to various AI services. Skip this if your environment explicitly forbids external authentication delegation.

How to get started

Access is granted through a standard OAuth authorization flow, delegating authentication externally to oauth2-proxy and Keycloak. The WANAKU_AUTH_ISSUER environment variable must be set to the Keycloak realm URL.

Use cases

• Exposing Apache Camel routes as AI tools, allowing AI agents to interact with enterprise systems without direct access.
• Multi-agent orchestration, where Wanaku coordinates multiple AI services, such as querying documents, generating responses, and handling tool execution.
• Functioning as an enterprise AI gateway to standardize AI access, enabling developers to use a single API for various underlying AI services like OpenAI, Anthropic, or vector databases.

Wanaku provides a crucial layer for integrating AI agents with enterprise infrastructure, offering a governed proxy for actions. Its support for dynamic tools, multiple protocols, and external authentication makes it adaptable for complex environments. It simplifies AI application development and access standardization.