Kiln

Kiln exposes both a command-line interface and an MCP server that lets AI agents orchestrate and execute physical fabrication jobs across multiple printers. Through the MCP surface, agents can discover printers on the network, authenticate printers of different types (OctoPrint, Moonraker/Prusa Link, Bambu, Elegoo), queue print jobs via the fulfillment layer (Craftcloud) or external provider integrations, and perform end-to-end print workflows from model search to slicing and printing. The MCP tooling provides a unified set of 353 MCP tools designed for AI agents, enabling tasks such as discovery, authentication, job creation, preflight checks, and print monitoring across a heterogeneous fleet. Kiln also supports external MCP-enabled agents, so you can route jobs between local printers and provider integrations without leaving the same workflow.

To use Kiln via MCP, run the Kiln MCP server (as shown in the installation steps) and connect your MCP-enabled agent to the server. Use the standard MCP tool invocations to perform actions like discovering printers, adding a printer, submitting a job, or querying status. Kiln’s CLI mirrors the MCP surface, so agents can perform tasks through either the CLI or MCP commands, enabling a single, consistent automation experience whether you’re operating locally or orchestrating through a fulfillment or provider network.

Prerequisites:

• A machine on the same LAN as your printers (or network access to printers)
• Python 3.8+ installed
• Optional: PrusaSlicer or OrcaSlicer for direct in-Kiln slicing, and OpenSCAD for Gemini Deep Think model generation

Step-by-step installation:

1. Clone the Kiln repository or install from an existing clone: git clone https://github.com/codeofaxel/Kiln.git cd Kiln

2. Install Kiln in editable mode (recommended for development or local runs): python -m pip install -e .

3. Set up environment (optional Gemini API key): export KILN_GEMINI_API_KEY=your-gemini-api-key

4. Run the Kiln MCP server (Python module entry point): python -m kiln

5. Verify installation by using Kiln CLI for quick tasks like discovery and authentication: kiln discover kiln auth --name my-printer --host http://octopi.local --type octoprint --api-key YOUR_KEY

6. If you prefer to run via a script or in a container, use the corresponding execution approach (for example, the npm or container-based flows if you adapt them for your environment).

Prerequisites recap:

• Printer network reachability (LAN)
• Printer-specific credentials (API keys or tokens as applicable)
• Optional slicing and Gemini generation tools if you plan to use those capabilities

Environment variables and configuration:

• KILN_GEMINI_API_KEY enables Gemini Deep Think for model-to-geometry generation; set this if you plan to generate printable models from text descriptions. Kiln will skip Gemini-related steps if the key is not provided.
• Ensure printers are reachable on the LAN and that any required API keys (OctoPrint, PrusaLink, Moonraker) are configured when authenticating printers.
• Kiln supports multiple printer types and providers; when mixing local printers with fulfillment or external integrations, Kiln will route jobs automatically based on availability and user preferences.
• For best results, keep PrusaSlicer or OrcaSlicer installed if you intend to slice STL files directly from Kiln. OpenSCAD can be used for AI-generated models when Gemini is enabled.
• If you encounter network or authentication issues, verify firewall rules and ensure the MCP server has access to all printer control endpoints on the LAN.