Skip to content

README.md

Latest commit

 

History

History
109 lines (83 loc) · 4.76 KB

README.md

File metadata and controls

109 lines (83 loc) · 4.76 KB
PrintMCP

📚 PrintMCP Documentation

Welcome! This is the full documentation for PrintMCP — which lets you 3D print just by talking to your AI assistant, taking a job from "I want to print a coffee cup" all the way to plastic on the bed.

You never operate PrintMCP directly. You chat with your assistant (Claude, or whatever client you've connected) and it does the finding, slicing, and printing for you. The tutorials below teach you how — in plain conversation, not commands. The reference pages go deeper for the curious.

🚦 Start here

New to PrintMCP? Follow this path in order:

  1. Getting Started — a one-time setup: install PrintMCP and connect it to your assistant. ~10 minutes.
  2. Tutorial 1 · Find Something to Print — ask your assistant to find and download a model.
  3. Tutorial 2 · Get It Print-Ready — describe how you want it printed, in plain words.
  4. Tutorial 3 · Send It to Your Printer — start and watch a real print, safely.
  5. Tutorial 4 · From Idea to Object — the whole thing in one conversation.

🗂️ All documentation

Guides

Page What it covers
Getting Started Install with uv, create .env, verify, register with a client.
Configuration Every environment variable, how to get each credential, and where files are stored.
Safety Model Why physical-actuation tools need confirm=true, and how the dry-run gate protects your machine.
Troubleshooting Common errors at each level and how to fix them.
Architecture How the one-server / three-level design fits together (for contributors).
Releasing How PrintMCP is published to PyPI (for maintainers).

Tool reference (under the hood)

You don't call these yourself — your assistant does. These pages are for the curious, or for developers building on PrintMCP. Start with the developer overview for the invocation model, schemas, annotations, and error contract that all tools share.

Page What your assistant uses it for
Developer overview The shared contract: params envelope, annotations, response formats, error handling, programmatic calls
Level 1 · Thingiverse Finding and downloading models
Level 2 · Cura Slicing models into print files
Level 3 · OctoPrint Uploading, printing, and controlling the printer

Tutorials — learn by chatting

These teach you what to say to your assistant, not commands to type.

# Tutorial What you'll do
1 Find Something to Print Ask for a model and get real options, with licenses checked.
2 Get It Print-Ready Describe the print you want — fast, strong, smooth — in plain words.
3 Send It to Your Printer Start, watch, and pause a real print — with a safety check on anything physical.
4 From Idea to Object Go from request to finished print in one conversation.

🧭 What happens when you ask

flowchart LR
    A(["🗣️ 'print a coffee cup'"]) --> L1
    subgraph L1["You describe what you want"]
        T["assistant finds a model"]
    end
    subgraph L2["You say how to print it"]
        C["assistant slices it"]
    end
    subgraph L3["You give the go-ahead"]
        O["assistant runs the printer"]
    end
    L1 --> L2 --> L3 --> Z(["✅ finished print"])
Loading

You only ever do the talking; your assistant handles the three stages — finding, slicing, and printing. (Each works on its own, too: you can find and download models even before your printer is set up. See Architecture for how it's built.)

💬 How to read the tutorials

  • Lines marked 💬 You are what you type to your assistant.
  • Lines marked 🤖 Assistant are roughly what it says back. (Exact wording varies — it's a conversation, not a script.)
  • You never type commands or file paths. If you see something technical, it's optional background for the curious.

Tip

Reading on GitHub? The Mermaid diagrams and > [!NOTE] callouts render automatically. In a plain text editor they'll appear as code/quote blocks — that's expected.