Skip to content

Getting Started with Wanaku

Welcome to Wanaku! In this guide you will install the Wanaku CLI, start a local instance, add a tool, and use it from an AI agent. By the end you will have a working MCP router that connects your agent to live data — no containers, no complex setup.

What You Will Need

  1. Java 21 or later
  2. An agent that supports the Model Context Protocol (MCP).

Step 1: Install the Wanaku CLI

The CLI ships as native binaries for Linux (x86_64) and macOS (AArch64), or as a Java-based archive that runs on any OS with Java 21+. Pick the tab that matches your system:

shell
curl -sSL https://wanaku.ai/get-wanaku.sh | bash
shell
wget https://github.com/wanaku-ai/wanaku/releases/download/v0.1.3/wanaku-cli-0.1.3-osx-aarch_64.zip
unzip wanaku-cli-0.1.3-osx-aarch_64.zip
install -m 750 wanaku-cli-0.1.3-osx-aarch_64/bin/wanaku wanaku-cli-0.1.3-osx-aarch_64/bin/wanaku-cli $HOME/bin/
rm -rf wanaku-cli-0.1.3-osx-aarch_64 wanaku-cli-0.1.3-osx-aarch_64.zip
shell
wget https://github.com/wanaku-ai/wanaku/releases/download/v0.1.3/wanaku-cli-0.1.3-linux-x86_64.zip
unzip wanaku-cli-0.1.3-linux-x86_64.zip
install -m 750 wanaku-cli-0.1.3-linux-x86_64/bin/wanaku wanaku-cli-0.1.3-linux-x86_64/bin/wanaku-cli $HOME/bin/
rm -rf wanaku-cli-0.1.3-linux-x86_64 wanaku-cli-0.1.3-linux-x86_64.zip
shell
# Requires Java 21+
wget https://github.com/wanaku-ai/wanaku/releases/download/v0.1.3/wanaku-cli-0.1.3.zip
unzip wanaku-cli-0.1.3.zip
# Add the extracted bin/ directory to your PATH

TIP

The Quick Install script auto-detects your OS and architecture, downloads the latest release, verifies the checksum, and installs to $HOME/bin. You can override the install directory with WANAKU_INSTALL_DIR=/usr/local/bin.

Verify the installation:

shell
wanaku --version

Step 2: Start Wanaku Locally

This single command downloads the router and the built-in capability services, then starts everything on your machine:

shell
wanaku start local

Authentication is disabled automatically in local mode, so there is nothing else to configure.

After a few seconds, open http://localhost:8080 in your browser. You should see the Wanaku UI.

TIP

You can enable additional capability services with the --services option. Run wanaku start local --help to see what is available.

Step 3: Check That Everything Is Running

Use the CLI to verify that the capability services registered with the router:

shell
wanaku capabilities list

You should see output similar to:

service serviceType  host      port status lastSeen
http    tool-invoker 127.0.0.1 9000 active Wed, May 28, 2026 at 10:00:00

This means Wanaku is up and running, and the HTTP capability service is ready to handle tool calls.

Step 4: Add a Tool

Wanaku uses toolsets to group related tools together. Let's import a currency conversion toolset so your agent can look up exchange rates.

Option A: Via the Web UI

  1. Open http://localhost:8080/admin/#/service-catalog
  2. Click Toolsets Repositories Tab
  3. Expand the wanaku-toolsets one.
  4. Import the currency one.

Option B: Via the CLI

Use this if you prefer to do everything via CLI.

shell
wanaku tools import https://raw.githubusercontent.com/wanaku-ai/wanaku-toolsets/refs/heads/main/toolsets/currency.json

Confirm the tool was added:

shell
wanaku tools list

Expected output:

name                          namespace type uri                                                                                                       labels 
free-currency-conversion-tool default   http https://economia.awesomeapi.com.br/last/{parameter.value('fromCurrency')}-{parameter.value('toCurrency')} {}

Step 5: Connect an AI Agent

Wanaku speaks the Model Context Protocol (MCP) over HTTP. Any MCP-compatible client can connect to it. How you do that will depend on the AI agent you are using.

  1. Claude Desktop: you can run Wanaku's command wanaku configure claude
  2. Cursor: you can run Wanaku's command wanaku configure cursor
  3. Claude Code: claude mcp add wanaku --transport sse http://localhost:8080/mcp/sse
  4. IBM Bob: details here.
  5. For any other client, http://localhost:8080/mcp/sse for SSE or http://localhost:8080/mcp/ for Streamable HTTP.

Step 6: Testing

Then ask it a question like:

How much is 100 USD in CZK?

Claude call example

What's Next?

You now have a working Wanaku instance with a real tool. Next, learn how to manage tools, resources, prompts, and forwards:

If something does not work, check the logs printed by wanaku start local in your terminal.

If you find a bug, please report it. To get in touch with the community, visit the Wanaku project.