Use an agent to build a full engineering package
A developer used Dots to iterate on a large mechanical and electrical design until it produced CAD, calcs, drawings, and a BOM.
- What they did
- They handed Dots a sizable engineering job, answered a few follow-up questions, and then let it work for hours or days on a cloud machine. In some cases, they also dropped a task brief into Google Drive and let the agent work through the items in sequence. The setup included CAD tools, schematic software, Blender, and scheduling, so it could review files, edit models or code, run simulations, and compare the results before changing the design again.
- How it went
- They said one package ended up with more than 1,000 unique parts, over 6,000 solids, close to 300 pages of drawings, and hundreds of electrical connections. The system also shrank a 200k-triangle model through several lighter versions, and a set of laser-cut parts worked after a small speed adjustment. They noted the first structural pass used about a 5x safety margin, and said they would not trust safety-critical work without human review.
- Tips
- Plan for long runtimes and check the economics before relying on it heavily. If the design is overbuilt at first, reset it with tighter requirements and review outputs carefully before manufacturing.
Try it yourself with Dots
Create a complete engineering package for [system name] with these requirements: [loads, dimensions, standards, interfaces, and constraints]. Work through design, calculation, and verification loops on your own until the results are consistent. Produce CAD files, calculation notes, a BOM, [electrical documents if needed], and manufacturing drawings in [folder]. Before you send anything for manufacture, pause and ask me to approve the final package.
Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.
Get 3 agent uses like this every morning
With the day's top AI news. Free, and one click to unsubscribe.
Related agent uses
Map a desktop app’s platform layers for Linux support
They used Claude to inspect a minified desktop app and trace why it depended on macOS, then outlined three possible Linux paths.
Repair a formal proof by iterating on code and builds
Codex fixed CertiGC proofs in Rocq with human review on any change that could alter the specification.
Use Codex to move research work faster
A theoretical chemist divided research tasks across models to speed up simulation-heavy work, while keeping a final check for tougher decisions.
Search for room-temperature magnetic semiconductor candidates
A Claude Opus 5.5 agent team screened Luttinger-compensated magnets and surfaced two predicted room-temperature magnetic semiconductor candidates.
Validate a circuit model with scope measurements
Claude Code can compare a SPICE run with oscilloscope readings from the real circuit.
Calibrate a robot hand and test gestures safely
Use an OpenClaw agent to inspect a 16-joint printed hand, verify motion with a camera, and calibrate firmware before trying simple gestures.