展示HN:我用AI代理将草图变成了一个为孩子准备的3D打印洞洞板。
Show HN: I turned a sketch into a 3D-print pegboard for my kid with an AI agent

原始链接: https://github.com/virpo/pegboard

该项目详细介绍了根据简单的手绘草图创建的可定制、3D打印的洞洞板玩具的过程。利用人工智能(Codex)和Python脚本,创作者在Fusion 360中快速生成和迭代设计,与传统CAD方法相比,大大缩短了设计时间。 与手动建模每个零件不同,参数化生成器允许轻松修改和扩展——甚至可以由人工智能代理完成。当前套装包括七个游戏零件、四个齿轮和两个洞洞板,所有都基于40毫米网格系统。 该仓库不仅提供最终的STL文件,还提供用于生成这些文件的Python脚本,以及清晰的说明(“AGENTS.md”),说明如何使用编码代理扩展该套装,添加新的零件、尺寸或功能。这使得玩具具有高度适应性,并鼓励进一步的创造性探索。重点是快速的设计-打印-测试循环,优先考虑游戏和定制。

一位 Hacker News 用户名为 virpo,分享了一个项目,他使用 AI 代理(Codex)将简单的草图快速转化为可 3D 打印的洞洞板,给孩子使用。 他没有手动在 Fusion 360 中设计洞洞板——这个过程需要几个小时——而是简单地向 AI 提供了草图照片和关键尺寸(孔间距和插销宽度)。 在五分钟内,AI 生成了一个可打印的设计,经过一些小的调整后,效果完美。Virpo 已经在 GitHub 上分享了这个项目,供其他人使用和修改。他认为这是一个有趣且“奇怪”的 AI 应用,强调了即使在时间有限的情况下,AI 也能释放创造力,另一位评论者也对此表示赞同,称赞了 LLM 空间中的“代理 x 家长”组合。
相关文章

原文

AI-generated 3D-printable pegboard toy from a hand-drawn sketch.

We have pegboards and plywood scraps all over the apartment. I wanted to make a tiny one for Oli, or maybe even hang one on our front door instead of a wreath so visitors would have something to play with.

I had already cut the scrap wood and drilled the first board when I sat down at the computer, ready to lose an hour or two in Fusion 360. Then I looked at the sketch on my desk, took a photo, pasted it into Codex, and added the only dimensions that mattered: the holes are 40 mm apart and the pegs are 8 mm wide.

About a minute later I had the first set of pieces. From there I just iterated a little: print, test, adjust, repeat, until the pegs fit snugly, the pieces felt good by hand, and the gears turned smoothly.

Everything in this repo stays as small Python generators instead of hand-edited meshes, which made those tweaks easy. The time I did not spend drawing every variant in CAD turned into time I could spend printing, testing, and playing with Oli instead.

The current set is a 40 mm system with seven play pieces, one tuned peg, four gears, and two printable boards.

This repo is intentionally easy to modify with coding agents. AGENTS.md has the current dimensions, folder layout, and the few rules needed to extend the set safely.

You can ask an agent to:

  • build a bigger pegboard like 6x6
  • make the pegs longer or shorter
  • add new pieces for different peg combinations
  • scale the whole system up or down
  • generate tighter or looser fit-test variants

What's in this repository

Overview render showing the play pieces, gears, printable pegboards, and tuned pegs included in the repository

Play pieces Gears Pegboards and pegs
Seven flat pieces in models/pieces/ with 8.45 mm holes, tuned to lift on and off the pegs easily. Four smooth gears in models/gears/, tuned to mesh on the 40 mm peg grid. Two printable boards in models/boards/ plus the tuned peg in models/pieces/.
  • models/
    Final STL exports plus the board-fit prototypes.
  • scripts/
    Parametric generators for pieces, gears, boards, and README assets.
  • docs/
    Supporting notes plus the README images.
  • AGENTS.md
    Instructions for coding agents that want to extend the set.
  • Grid pitch: 40.0 mm
  • Piece hole diameter: 8.45 mm
  • Gear hole diameter: 8.45 mm
  • Peg: 7.72 mm diameter, 40.0 mm long, 1.2 mm end roundover
  • Printed board hole diameter: currently 8.30 mm, still being validated
python3 -m venv .venv
. .venv/bin/activate
pip install -r requirements.txt
python scripts/generate_pegboard_shapes.py
python scripts/generate_pegboard_gears.py
python scripts/generate_pegboard_board.py
python scripts/generate_repository_assets.py
联系我们 contact @ memedata.com