Skip to content

Quick Start Guide

Emmu is designed with dual interfaces: a Terminal-First Interactive CLI for direct agentic chat, and a Sleek Web UI for managing persistent campaigns and reviewing technical findings.

[!NOTE] This guide assumes you have completed the Installation and Configuration steps.


1. Launching the Platform

The entry point for Emmu is run.py. Depending on the flags you pass, Emmu will boot into different operational modes.

Interactive CLI Mode (Default)

To start a persistent terminal chat session with the Synchronous Orchestrator:

python run.py
This drops you into a conversational interface where you can instruct Emmu to perform reconnaissance, reverse engineer a file, or analyze a domain.

One-Shot Execution

For automation scripts or cron jobs, you can pass a single prompt directly:

python run.py --prompt "Run a fast port scan on example.com and summarize the results."

Web UI + API Server Mode

To launch the background API and the Campaign Dashboard (emmuOEC, Target Manager):

python run.py --web
Access the dashboard by navigating to http://localhost:5000 in your browser.


2. Essential CLI Commands

When inside the Interactive CLI chat, you can use Slash Commands (/) to control the framework natively.

Command Action
/help Display the list of available commands.
/status View the status of your connected MCP servers and loaded Specialists.
/model Switch the active LLM provider (e.g., from Groq to OpenRouter).
/web start Launch the Web Dashboard from within the chat session.
/web stop Terminate the Web Dashboard background process.

3. Your First Mission

Let's test Emmu's reasoning and tool-calling capabilities. From the Interactive CLI, issue the following command:

Initialize a workspace for "Target Alpha". Discover the subdomains for hackerone.com and save the top 5 most interesting results to my scratchpad.

What happens next? 1. Emmu evaluates the prompt and selects the Recon Specialist. 2. Emmu uses the init_mission_workspace tool to create persistent storage. 3. It utilizes an MCP tool (like subfinder or a generic linux command) to enumerate domains. 4. The Reviewer Agent verifies the output. 5. Emmu writes the findings using the emmu_update_scratchpad tool.


Next Steps

To truly understand how Emmu achieves this without hallucinating, dive into the architecture starting with The Synchronous Orchestrator.