Connect Spexus to your AI agent

Step 1: Download Spexus MCP

Download the archive from the releases page and unpack the binary into a convenient directory such as `/usr/local/bin/`.

tar -xzf rms-ai-darwin-arm64.tar.gz
mv spexus-mcp /usr/local/bin/spexus-mcp

Step 2: Initialize the client

On first launch, run the initialization and provide the server address `https://app.spexus.ai`, your username, and your password.

/usr/local/bin/spexus-mcp -i

Step 3: Connect it to your AI agent

claude mcp add spexus /usr/local/bin/spexus-mcp

codex mcp add spexus /usr/local/bin/spexus-mcp

opencode mcp add

{
  "mcpServers": {
    "spexus": {
      "command": "/usr/local/bin/spexus-mcp",
      "args": [],
      "disabled": false,
      "autoApprove": []
    }
  }
}

Step 4: Verify the connection

Launch your AI agent and ask: "Show the epics from Spexus." If everything is configured correctly, the agent will call the MCP server and return the list of epics.

Step 5: Give the agent its first task

Now you can immediately use the connection in real work. Ask the agent: "Create a new epic on the topic of ..." and then describe what you want to build, what result you need, and which context matters.

Create a new epic on the topic of ...

If MCP is configured correctly, the agent can create a new epic directly in Spexus and start refining the requirements in dialogue.