dsh-print3d
Run the following command in DeepSeek Harness:
dsh plugin install weisofns/dsh-print3d
Paste the following prompt into your AI chat to install this plugin:
Exit DSH completely, then run dsh plugin install weisofns/dsh-print3d; the full source is at https://github.com/weisofns/dsh-print3d
About this plugin
Running a full 3D-printing workflow inside an agent conversation—parametric modeling, calibration G-code, slicing estimates, toolpath inspection—normally means hand-cranking scripts, flipping through material parameter sheets, and stuffing megabytes of G-code text into the context window until a 7B model loses the thread. dsh-print3d turns every step into six first-class agent tools and five on-demand knowledge skills. Generated STL, G-code, and PNG files are written straight to a local output folder and only the file path is returned to the model, keeping the context lean enough for small local models to stay on task end to end.
The tool layer covers parametric STL generation (box, cylinder, tube, sphere, cone, rounded box, and gear), five categories of calibration-part G-code (cube, temp-tower, first-layer, retraction, bridge), STL watertightness and overhang analysis, G-code print-time and material estimation, toolpath top-down PNG rendering, and an optional bridge to a local PrusaSlicer install. The skill layer structures engineering knowledge—twelve material parameter tables, Marlin vs. Klipper dialect differences, STL and G-code metric interpretation, and a ten-fault diagnostic matrix—so the agent loads exactly the reference it needs via tool-skill without bloating the system prompt.
The plugin is preset-agnostic: any DSH agent, regardless of persona, can call the print3d tools and load the print3d skills out of the box with zero extra npm dependencies. It is especially well suited to users running local small models such as Ollama + Qwen3 8B, where the bundled concise persona is tuned for short-instruction prompting and the entire model, estimate, and diagnose cycle runs offline with no API cost.
Use Cases
- Go from parametric STL generation to G-code print-time estimation in one agent conversation
- Run the full model, calibrate, preview, and diagnose loop offline with Ollama + Qwen3 8B
- Any preset agent can call print3d tools out of the box with zero extra dependencies
Best For
- Solo engineers running local LLMs for 3D-printing workflows
- Teams sharing a reusable 3D-printing toolchain across multiple agents
- Prompt designers who prefer file-path returns over context-heavy text dumps
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.