Wanaku — A Governed Action Proxy for AI Agents

Documentation · Releases · Discussions

Wanaku is a governed action proxy for AI agents. It sits between agents and the systems they act on, intercepting Model Context Protocol (MCP) tool calls, agent-to-agent messages, and inference traffic. Before version 0.2.0, the project was named Wanaku MCP Router.

Wanaku supports open standards for AI agents and provides first-class integration with enterprise frameworks, including Apache Camel and Quarkus. For example, integration developers build Apache Camel routes and use Wanaku's Camel Integration Capability to publish them as tools. Agents call those tools with parameters, but Wanaku runs the work. Agents never access backend systems directly. The proxy enforces policy, identity, data controls, and audit requirements.

The project name comes from the origin of guanaco, a camelid native to South America.

Key Features

• Agent Isolation — Agents call tools through Wanaku; they never reach backend systems directly

• Policy Enforcement — LLM-powered evaluators + WASM action scripts classify, filter, and block tool calls in the proxy layer

• Identity & Auth — Authentication and authorization via oauth2-proxy and Keycloak, enforced before actions reach backends

• Tool Discovery — Auto-discover tools from upstream MCP servers; integration developers publish Camel routes as tools

• Namespace Isolation — Organize tools and resources across isolated namespaces per team, tenant, or environment

• Extensible Architecture — Feature crates, a composable filter pipeline, and Admin UI plugins

• Admin Dashboard — Web UI for managing tools, resources, prompts, and forwards

• Container-Ready — Multi-arch images (x86_64, aarch64) published automatically

Related MCP server: mcp-mesh

Quick Start

Install

On Linux or macOS, use this command to install the latest early-access build. Early-access builds are development builds of the unreleased Wanaku 0.3.0.

curl -fsSL https://raw.githubusercontent.com/wanaku-ai/wanaku/main/get-wanaku.sh | bash

The installer puts wanaku-server in $HOME/bin.

This README describes the Rust-based Wanaku server. The releases up to v0.2.0, including the release in the badge above, contain the Java-based Wanaku, which is deprecated. To move from it, read Migrate from Classic Wanaku.

Other install methods are available. You can download the archive manually, use a container image, run Wanaku as a systemd service, or deploy it on Kubernetes or OpenShift. Read the Quickstart for the details of each method.

Run wanaku-server. Open http://localhost:8080 to see the dashboard:

If you prefer a text interface, you may also download and use the Wanaku CLI from the Wanaku Barn project.

Documentation

The full documentation for this version is on the Wanaku website. The same pages are in the docs/ folder of this repository:

• Get started: install Wanaku, run your first governed MCP proxy, and learn the architecture.

• Deploy and operate: run Wanaku on Kubernetes, persist state, add authentication, and secure the deployment.

• Connect agents and backends: proxy A2A traffic, inference requests, and semantic routers.

• Govern agent actions: control agent actions with posture, action policies, evaluators, credentials, and audit.

• Extend Wanaku: write evaluator actions, feature crates, and Admin UI plugins.

• Reference: environment variables, configuration files, ports, and the management API.

To learn Wanaku step by step, follow the guided tutorial. To contribute, read CONTRIBUTING.md. To report a vulnerability, read SECURITY.md.

Community

• GitHub Issues - Bug reports and feature requests

• Discussions - Ask questions and share ideas

Related Projects

• Wanaku Barn — Utilities for the Wanaku project: OpenShift/Kubernetes operator, CLI client, etc

• Camel Integration Capability — Build Apache Camel routes and publish them as tools that agents can call through Wanaku

License

This project is licensed under the Apache 2.0 License - see the LICENSE file for details.