Skip to main content
Glama
cadlens-mcp

cadlens-mcp

Official
by cadlens-co

cadlens-mcp

npm version GitHub

A Model Context Protocol server that wraps the Cadlens CAD parsing API so MCP-aware LLM clients (Claude Desktop, Claude Code, Cursor, Zed, Windsurf) can parse CAD files (.dwg, .dxf, .dwf, .dwfx, .dgn V7, .pdf, max 100 MB) and reason over the extracted entity, layer, and metadata payloads.

Cadlens converts CAD drawings into structured JSON without requiring AutoCAD or any desktop software — learn more at cadlens.co.

Install

Get an API key from the Cadlens dashboard first — keys start with cadl_ and are created in the dashboard for free.

Claude Desktop / Cursor / Windsurf

Add to your MCP client config (e.g. ~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "cadlens": {
      "command": "npx",
      "args": ["-y", "@cadlens/mcp-server"],
      "env": {
        "CADLENS_API_KEY": "cadl_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
      }
    }
  }
}

Claude Code CLI

claude mcp add cadlens \
  --env CADLENS_API_KEY=cadl_xxx \
  -- npx -y @cadlens/mcp-server

Related MCP server: DWG MCP Server

Development (build from source)

npm install
npm run build

export CADLENS_API_KEY="cadl_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
node dist/index.js

Tools

Tool

What it does

cadlens_parse_file

Upload a local CAD file, poll until parsed (5 min budget), return summary.

cadlens_parse_url

Download a CAD file from a URL, then parse it like parse_file.

cadlens_get_job

Cheap status check (PENDING / PROCESSING / COMPLETED / FAILED).

cadlens_get_result

Fetch parsed content. mode: summary (default), entities_by_type, entities_on_layer, full.

cadlens_refresh_image_url

Re-fetch the 1h presigned PNG URL without re-downloading the full result.

cadlens_list_jobs

The 100 most recent jobs for the configured API key.

cadlens_delete_job

Delete a job and its S3 artifacts. Irreversible.

Configuration

Env var

Required

Default

Notes

CADLENS_API_KEY

yes

—

Created in the CADLens dashboard.

CADLENS_API_BASE

no

https://api.cadlens.co/v1

Set to http://localhost:3001/v1 for local dev.

WEBHOOK_PORT

no

0 (random)

Port for the in-process webhook receiver.

WEBHOOK_PUBLIC_URL

no

unset

Set to a tunnel URL (ngrok/cloudflared) to let CADLens hit the local receiver. When set, parse calls auto-register the webhook and the poller short-circuits on receipt.

REQUEST_TIMEOUT_MS

no

30000

Per-HTTP-request timeout for CADLens calls.

Webhook short-circuit (optional)

If WEBHOOK_PUBLIC_URL is set, parse_file / parse_url register a per-process webhook URL alongside the upload. The webhook handler updates an in-memory job-state cache; the poller checks that cache before each HTTP GET and returns early when COMPLETED / FAILED arrives. This trims worst-case latency by up to one full poll tick (~1 s) without changing the tool surface.

Example tunnel setup:

cloudflared tunnel --url http://localhost:8787 &
export WEBHOOK_PORT=8787
export WEBHOOK_PUBLIC_URL="https://<your-tunnel>.trycloudflare.com"

Development

npm run typecheck
npm run lint
npm test
npm run smoke   # tools/list smoke test against built binary

Project-scoped Claude agents

This repo ships five agents under .claude/agents/:

  • cadlens-api-debugger — diagnoses unexplained CADLens 4xx/5xx using mcp-server-reference.md.

  • mcp-tool-tester — drives JSON-RPC against the built server to validate tool responses.

  • mcp-tool-implementer — scaffolds new tools following the existing src/tools/* pattern.

  • cad-drawing-summarizer — uses the MCP tools to summarize a CAD file in natural language.

  • cad-layer-inspector — drills into a single layer of a parsed drawing.


GitHub Topics

Add these topics to this repo for discovery: mcp mcp-server model-context-protocol ai-agents claude cad dwg dxf cad-api llm-tools engineering-api


License

MIT

Available Tools

7 tools
cadlens_delete_jobA

Delete a job and its uploaded file and preview image from CADLens storage. Irreversible.

Inputs: job_id (required).

cadlens_get_jobA

Check the status of a CAD parse job. Returns PENDING, PROCESSING, COMPLETED, or FAILED.

Inputs: job_id (required).

cadlens_get_resultA

Fetch the parsed result of a completed CAD job. Use mode='summary' (default) for an overview, 'entities_by_type' or 'entities_on_layer' for filtered detail, or 'full' for the complete vector JSON (can be very large — only use when explicitly needed).

Inputs: job_id (required), mode, layer_name, entity_type.

cadlens_list_jobsA

List the 100 most recent CAD parse jobs for the configured API key, newest first.

Inputs: no named parameters.

cadlens_parse_fileA

Parse a local CAD file (DWG/DXF/DWF/DWFx/DGN-V7/PDF, max 100 MB). Polls until the job completes or 5 minutes elapse. Returns a summary including the preview image URL. Use cadlens_get_result with the returned job_id to fetch detailed entity data.

Inputs: path (required), webhook_url, notify_email.

cadlens_parse_urlA

Download a CAD file from a URL (max 100 MB) and parse it. Same output as cadlens_parse_file.

Inputs: url (required), webhook_url, notify_email.

cadlens_refresh_image_urlA

Re-fetch the presigned preview-image URL for a completed job. Cheaper than re-fetching the full result. The URL is valid for 3600 seconds.

Inputs: job_id (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A4.2/5.0

    Scored across 7 tools

    Disambiguation5/5

    Each tool has a distinct purpose: two parse variants clearly separated by input source (local file vs URL), plus status checking, result retrieval, image refresh, listing, and deletion. There is no functional overlap between any tools.

    Naming Consistency5/5

    All tools follow the consistent pattern 'cadlens_<verb>_<object>' with snake_case (e.g., parse_file, get_result, delete_job). The naming is uniform, predictable, and clearly indicates both the action and the resource.

    Tool Count5/5

    Seven tools is an ideal size for a CAD parsing service. Each tool covers a necessary part of the workflow without redundancy or bloat, making the toolset well-scoped and easy to navigate.

    Completeness5/5

    The toolset covers the full lifecycle of a parse job: creation (from file or URL), status polling, result retrieval with multiple detail modes, preview image refresh, job listing, and deletion. No critical operations are missing.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • F
      license
      Not graded
      quality
      D
      maintenance
      MCP server that exposes CAD geometry reasoning over STEP files to LLMs, allowing natural language queries about parts, assemblies, dimensions, holes, and mass properties.
      -
    • A
      license
      Not graded
      quality
      B
      maintenance
      An MCP server for read-only inspection of AutoCAD DWG files, enabling AI agents to open drawings, query objects by handle or filter, and explore properties and references.
      22
      GPL 3.0
    • A
      license
      A
      quality
      D
      maintenance
      MCP server for AutoCAD LT automation and headless DXF generation, exposing tools for drawing, entities, layers, blocks, annotations, P&ID, and view operations via natural language.
      8
      MIT