Vectorize — MCP setup
MCP serverDocs & knowledgeHosted MCP front door to the Vectorize API: raster-to-vector tracing to SVG, PDF, EPS or DXF.
Available today. Use it from your connected AI after setup.
Needs your own Vectorize — MCP setup account. You sign in to it and approve access when you connect.
Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.
Then ask your AI: use Vectorize
Install Vectorize — MCP setup
The server’s own address, for the clients that take one directly. Or connect ahel once and every client you use reads it from one address, with the account kept on ahel rather than in each client’s config.
Claude Code
claude mcp add --transport http --scope user vectorize-mcp-setup 'https://vectorize.dudko.dev/mcp'Run it once in your project, then open /mcp to approve any sign-in the server asks for.
Claude Desktop
https://vectorize.dudko.dev/mcpAdd a custom connector in Settings, paste this address, and approve the sign-in.
Cursor
cursor://anysphere.cursor-deeplink/mcp/install?name=vectorize-mcp-setup&config=eyJ1cmwiOiJodHRwczovL3ZlY3Rvcml6ZS5kdWRrby5kZXYvbWNwIn0=Open the link and Cursor adds the server at that address.
ChatGPT
https://vectorize.dudko.dev/mcpIn Settings, enable Developer mode, create an MCP app, and paste this address. Your plan and workspace must allow custom apps.
Codex
codex mcp add vectorize-mcp-setup --url 'https://vectorize.dudko.dev/mcp'Run it once, then sign in with codex mcp login vectorize-mcp-setup if the server asks for an account.
From the project's README
As published by dudko-dev/vectorize-mcp-setup in README.md.
Connect your AI assistant to Vectorize — raster images (PNG, JPG, …) traced into vector drawings: SVG, PDF, EPS or DXF — through our hosted Model Context Protocol (MCP) server. Sign in once with your dudko.dev account; no API key to paste.
The server is the front door to the Vectorize HTTP API, not a tracer: a picture never travels through the model or through the server. The assistant asks for a request, and runs it wherever the image is.
- Server:
vectorize - Transport: streamable HTTP
- URL:
https://vectorize.dudko.dev/mcp - Auth: OAuth 2.1 via
auth.dudko.dev— you'll be prompted to sign in.
What you can do
Three tools, all free:
| Tool | What it does |
|---|---|
api_access | How the Vectorize API works. |
trace_options | The presets (photo, poster, black & white), the options and the limits. |
api_request | Checks trace options and returns the exact request — URL, headers, a curl command — with a 15-minute credential for your account. Output: svg, pdf, eps, dxf or json. |
Install in Claude Code (plugin)
/plugin marketplace add dudko-dev/vectorize-mcp-setup
/plugin install vectorize@vectorize
/reload-plugins
Then run /mcp and complete the sign-in when prompted.
Prefer not to use the marketplace? Add the server directly:
claude mcp add --transport http vectorize https://vectorize.dudko.dev/mcp
Connect from other clients
Same remote URL, same sign-in — only the config location differs.
Claude Desktop / claude.ai connectors
- Open Settings → Connectors → Add custom connector.
- Name:
Vectorize - URL:
https://vectorize.dudko.dev/mcp - Save, then connect and sign in when prompted.
Cursor
~/.cursor/mcp.json (global) or .cursor/mcp.json (per project):
{
"mcpServers": {
"vectorize": {
"url": "https://vectorize.dudko.dev/mcp"
}
}
}
VS Code (GitHub Copilot / MCP) and other clients
{
"servers": {
"vectorize": {
"type": "http",
"url": "https://vectorize.dudko.dev/mcp"
}
}
}
Authentication
The server is an OAuth 2.1 resource server. The first call answers 401 with a
WWW-Authenticate header pointing at its protected-resource metadata:
https://vectorize.dudko.dev/.well-known/oauth-protected-resource/mcp
which names https://auth.dudko.dev as the authorization server (scope
usage). Compliant MCP clients discover this automatically — you only need the
server URL, and you sign in with your dudko.dev account. Tokens are
audience-bound to https://vectorize.dudko.dev/mcp and are refused anywhere else.
The server itself carries no files and no content: its tools return a ready-to-run HTTP request with a short-lived credential (15 minutes, good for this API only) for the account you signed in as. Your assistant runs that request wherever the file is, and the API answers the caller directly. Each request counts against your account's plan like any other API call; the MCP tools themselves are free.
Troubleshooting
401/ asked to sign in again — the token expired; reconnect in the client.- "plan used up" /
402— the tracing quota for your account is spent; retrying will not help. See https://auth.dudko.dev/account. - Tools don't appear — check the client finished the sign-in (
/mcpin Claude Code shows connected).
Support
- Website: https://vectorize.dudko.dev
- API docs: https://vectorize.dudko.dev/api
- For agents: https://vectorize.dudko.dev/llms.txt
- Contact: sergey@dudko.dev
Signals
- Last commit
- Sep 2026
Advanced
- Delivery
- vectorize MCP server → your ahel connector (mcp.ahel.ai) → your AI.
- Item type
- mcp-server
- Key
dev-dudko-vectorize- Source
- github.com/dudko-dev/vectorize-mcp-setup
- Hosted endpoint
https://vectorize.dudko.dev/mcp
github.com/dudko-dev/vectorize-mcp-setup
More in Docs & knowledge
MCP server · builderio
More in Docs & knowledgeemailmd
MCP server · anypost
More in Docs & knowledgeknowl
MCP server · dat999zx
More in Docs & knowledgemcp-memory-server
MCP server · mnemoverse
More in Docs & knowledgehoncho
MCP server · plastic-labs
More in Docs & knowledgeagent-native-clips
MCP server · builderio
More in Docs & knowledge