🎩 You're Invited:Meet the Socket team at Black Hat in Las Vegas, August 3-6.RSVP
Sign In

techdrawai-mcp

Package Overview
Dependencies
Maintainers
1
Versions
2
Alerts
File Explorer

Advanced tools

Socket logo

Install Socket

Detect and block malicious and high-risk dependencies

Install

techdrawai-mcp

TechDraw AI: turn a product photo into a dimensioned 2D technical drawing, a CAD-ready DXF or DWG, a 3D mesh or a photoreal render, from your terminal or your AI agent. MCP server and CLI.

latest
Source
npmnpm
Version
0.1.1
Version published
Maintainers
1
Created
Source

techdrawai

Turn a photo of a physical object into a dimensioned 2D technical drawing, CAD-ready DXF/DWG geometry, a 3D mesh, or a photorealistic render — from your terminal or from an AI agent.

This package is both:

  • an MCP server, so Claude (or any MCP client) can draft, vectorise, reconstruct and render as tools; and
  • a CLI, so the same things work from a shell script or a build pipeline.

Both are thin clients over the TechDraw AI public API. Nothing is computed locally, so there is no CAD kernel to install.

Install

npm install -g techdrawai-mcp

Or run it without installing: npx techdrawai-mcp.

Sign in

You need a TechDraw AI account on the Studio plan — programmatic access is a Studio feature. Then:

techdrawai auth login

That opens your browser, creates an API key from your session, and hands it back to the terminal over loopback — you never copy or paste a key. It is stored at ~/.techdrawai/config.json with owner-only permissions.

If you would rather manage the key yourself, create one in Settings → API keys at https://techdrawai.com/app?view=settings and either run techdrawai auth tda_sk_… or set TECHDRAWAI_API_KEY.

techdrawai whoami     # who the stored key belongs to, verified against the server
techdrawai credits    # balance, plan and per-operation costs
techdrawai auth logout

Use it from an AI agent (MCP)

Add the server to your MCP client. For Claude Code:

claude mcp add techdrawai -- npx -y techdrawai-mcp

Or, in a client that reads a JSON config:

{
  "mcpServers": {
    "techdrawai": {
      "command": "npx",
      "args": ["-y", "techdrawai-mcp"],
      "env": { "TECHDRAWAI_API_KEY": "tda_sk_…" }
    }
  }
}

The env block is optional — with no key in the environment the server uses the one techdrawai auth login stored.

Tools

ToolWhat it doesCost
generate_technical_drawingPhoto → dimensioned 2D technical drawing3 credits
render_productTechnical drawing → photorealistic product render3 credits
image_to_dxfAny image → DXF (and DWG) CAD geometryfree
photo_to_3dPhoto → textured 3D mesh, GLB/FBX/OBJ/USDZ/STL3 credits
identify_productName the object in a photofree
check_creditsBalance, plan, and what each operation costsfree

Every tool writes its result to disk (./techdrawai-output/ unless you pass output_path) and returns the path. That keeps megabyte-scale images out of the agent's context, and makes chaining natural: the drawing a tool just wrote is the input to the next one.

A typical agent run:

Take bracket.jpg, work out what it is, draw it to ISO A3 with the overall width at 120 mm, then give me a DXF at that real size.

which becomes identify_productgenerate_technical_drawingimage_to_dxf.

Use it from a shell

# What is this thing?
techdrawai identify -i bracket.jpg

# Photo -> technical drawing, anchored to two real measurements
techdrawai draw -i bracket.jpg \
  --name "steel wall bracket" \
  --measure 120:"overall width" \
  --measure 45:"height" \
  --units mm --standard ISO --paper A3 --type working \
  --views front,top,side \
  --out drawings/bracket.png

# Drawing -> CAD geometry at real size, plus AutoCAD's DWG
techdrawai dxf -i drawings/bracket.png --width-mm 120 --dwg --out cad/bracket.dxf

# Drawing -> photoreal product shot
techdrawai render -i drawings/bracket.png --style studio --material "brushed aluminium"

# Photo -> 3D mesh
techdrawai model3d -i bracket.jpg --detail high --out models/bracket.glb

techdrawai help prints every flag.

Measurements are what make it accurate

Without --measure, dimensions are the model's best estimate and are marked as nominal. With even one real measurement, every other feature is scaled proportionally from it, so the whole drawing becomes internally consistent and manufacturable. The syntax is value:where it was taken — the wording of the second half also tells the drafter whether it is a width, a height or a diagonal.

Credits and plans

OperationCredits
Technical drawing3
Product render3
Photo to 3D3
Image to DXF / DWGfree
Identify productfree

Failed generations are refunded automatically. API, CLI and MCP access requires an active Studio plan — the free tier's introductory credits and the Starter plan are for the web app.

Configuration

VariablePurpose
TECHDRAWAI_API_KEYYour key. Overrides the stored config, so CI and containers need no interactive login.
TECHDRAWAI_BASE_URLPoint the client at a different deployment. Defaults to https://techdrawai.com.

Config file: ~/.techdrawai/config.json.

Exit codes

0 success · 2 out of credits or a plan is required · 1 everything else.

Licence

MIT

Keywords

mcp

FAQs

Package last updated on 28 Jul 2026

Did you know?

Socket

Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.

Install

Related posts