Independent Open-Source Project Not affiliated with Cisco Systems, Inc.

Build Packet Tracer labs
from natural language.

MCP Packet Tracer is an open-source server that turns high-level lab descriptions into topology plans, PTBuilder scripts, IOS configs, and real-time deployment into Cisco Packet Tracer.

Clone the Repo See How It Works 
$ python -m src.packet_tracer_mcp
INFO  Packet Tracer MCP v0.4.0
INFO  Transport: streamable-http
INFO  http://127.0.0.1:39000/mcp

> pt_plan_topology(
    routers=3,
    pcs_per_lan=2,
    dhcp=True,
    routing="ospf",
    has_wan=True
  )

✓ Plan created — 3 routers, 6 PCs
✓ 3 LANs + 2 WAN links
✓ DHCP pools configured
✓ OSPF area 0 on all routers
22 MCP Tools
5 Resources
34 Tests Passing
11 Device Models

Core Capabilities

From intent to deployed topology.
All through MCP.

Natural Language Input

Describe a network lab in plain language. The MCP server translates intent into a full topology plan with devices, links, and addressing.

Validation & Auto-Fix

Plans are validated against real Packet Tracer constraints. The auto-fixer corrects cables, upgrades models, and reassigns ports automatically.

Live Deploy to PT

Not just files — deploy directly into a running Packet Tracer session via the HTTP bridge. Watch your topology appear in real-time.

Full Artifact Export

Generate PTBuilder scripts, IOS CLI configs, project files, and natural language explanations. Everything you need to reproduce or teach.

Who is it for?

Built for anyone who works
with Packet Tracer.

Networking Students

Generate repeatable LAN, WAN, and multi-router labs faster than building every Packet Tracer file by hand.

Instructors

Create demos with addressing, DHCP, static routing, or OSPF — with topology plans and CLI output for step-by-step classes.

AI/MCP Developers

A concrete MCP use case where natural language results in a visible network simulation, not just text output.

Automation Engineers

Use the tool catalog, plan JSON, and execution layer as a base for custom lab generators or agent workflows.

Ready to automate?

Start building Packet Tracer labs with AI today.

Clone the repo, start the MCP server, connect your AI agent, and generate your first topology in minutes.

Clone the Repo Getting Started Guide 
git clone https://github.com/Mats2208/MCP-Packet-Tracer.git
cd MCP-Packet-Tracer
pip install -e .
python -m src.packet_tracer_mcp