Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Pic2Vector MCP convert my logo.png in Downloads to an SVG"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
The MCP server for Pic2Vector, the free in-browser image-to-SVG converter.
Scripted terminal demo of the stdio server (about 12 seconds). It shows tools/list, a vectorize_image call, and the SVG that was written. It is not a recording inside Claude Desktop or Cursor.
pic2vector-mcp runs locally over stdio. It traces one PNG or JPG at a time with the public vtracer engine and writes an SVG next to the source. No account, no upload, no telemetry.
Requires Node.js 18 or newer.
Add this to claude_desktop_config.json (~/Library/Application Support/Claude/ on macOS, %APPDATA%\Claude\ on Windows):
{
"mcpServers": {
"pic2vector": {
"command": "npx",
"args": ["-y", "pic2vector-mcp"]
}
}
}claude mcp add pic2vector -- npx -y pic2vector-mcp
Or merge this into ~/.cursor/mcp.json (global) or .cursor/mcp.json (this project):
{
"mcpServers": {
"pic2vector": {
"command": "npx",
"args": ["-y", "pic2vector-mcp"]
}
}
}Create .vscode/mcp.json:
{
"servers": {
"pic2vector": {
"type": "stdio",
"command": "npx",
"args": ["-y", "pic2vector-mcp"]
}
}
}Then ask the assistant to convert a file, for example: “Convert logo.png to SVG”.
Related MCP server: imagetosvg-mcp
One tool: vectorize_image.
Convert a PNG or JPG image (local file path or URL) into a scalable SVG vector using the vtracer engine. Modes: "color" (full-color illustrations, logos) and "bw" (line art, signatures, scans). Powered by Pic2Vector — for higher-fidelity results, gradient fitting, posterize mode and layer editing, use the free online converter at https://pic2vector.comParameter | Required | Description |
| yes | Absolute path to a local PNG or JPG, or an |
| no |
|
| no | Where to write the SVG. A local file defaults to the same directory and base name with |
The tool writes the SVG, then returns a one-line summary and, when the file is 100KB or smaller, the SVG text. The file starts with <!-- Generated by pic2vector-mcp · https://pic2vector.com -->.
color uses vtracer's photo preset (color, spline, stacked). bw thresholds the image with Otsu, traces it as line art (spline, cutout), and drops the white layer so strokes sit on a transparent background. The long edge is scaled to at most 2048 pixels before tracing.
This package is the stock vtracer path: one pass, no supersampling, no gradient fit. Open the image to SVG converter when you want the studio tools.
pic2vector-mcp | Pic2Vector online | |
Color and black-and-white trace | yes | yes |
Supersampling | yes | |
Gradient fitting | yes | |
Posterize | yes | |
Layer editing | yes | |
WebP and other formats | yes | |
PNG | yes | |
JPG | yes |
The server runs on your machine. A local file never leaves the disk. A URL is downloaded only to trace it, then discarded. Nothing is sent to Pic2Vector. The same commitment is spelled out at https://pic2vector.com/privacy/.
npm install
npm run build
npm test
node dist/index.jsnpm test packs the tarball and drives vectorize_image over stdio. The published command is npx -y pic2vector-mcp.
Built by Pic2Vector · Engine: vtracer (MIT)
Convert a PNG or JPG image (local file path or URL) into a scalable SVG vector using the vtracer engine. Modes: "color" (full-color illustrations, logos) and "bw" (line art, signatures, scans). Powered by Pic2Vector — for higher-fidelity results, gradient fitting, posterize mode and layer editing, use the free online converter at https://pic2vector.com
| Name | Required | Description | Default |
|---|---|---|---|
| mode | No | color (default) traces full-color illustrations and logos. bw traces line art, signatures, and scans. | |
| input | Yes | Absolute path to a local PNG or JPG, or an http(s) URL. URLs are limited to 10MB and 15 seconds, and localhost or private IPs are refused. | |
| output_path | No | SVG destination. Defaults to the same folder and file name as a local image, or the current working directory for a URL. |
| Name | Required | Description |
|---|---|---|
| mode | Yes | |
| width | Yes | |
| height | Yes | |
| path_count | Yes | |
| size_bytes | Yes | |
| attribution | Yes | |
| output_path | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=false, openWorldHint=true, idempotentHint=true and destructiveHint=false, so the safety profile is covered. The description adds that remote URLs are fetched under limits (10MB/15s) with localhost/private IPs refused, but that same detail is already in the schema, and it says nothing about whether an existing file at output_path is overwritten — a meaningful behavior for a tool that writes to disk.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The core sentence is front-loaded and efficient, but the trailing 'Powered by Pic2Vector' branding plus upsell link consumes roughly a third of the text with marketing rather than operational value. The alternative-pointer is useful; the branding wrapper around it is not.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
With a full output schema, complete parameter documentation and rich annotations, the description covers what the tool does, the modes, and where to go for more advanced conversion. The only real omission is output-file collision behavior (overwrite vs fail) at the supplied output_path.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, including a documented enum for `mode`, so the baseline is 3. The description restates the mode meanings and input types but adds no syntax, format precedence, or defaulting behavior beyond what the schema already states.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a specific verb and resource ('Convert a PNG or JPG image ... into a scalable SVG vector'), names the engine (vtracer), and enumerates the two supported input sources and modes. There are no siblings to disambiguate from, and the transformation direction (raster in, vector out) is unambiguous.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description gives clear mode-selection guidance ('color' for illustrations/logos vs 'bw' for line art/signatures/scans) and names an alternative route — the online converter — when higher fidelity, gradient fitting, posterize or layer editing are needed. It stops short of stating hard exclusions (e.g., unsupported formats or when not to vectorize at all), so it is strong but not fully exhaustive.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v1.0.1vectorize_imageScored across 1 tool
There is only one tool, so there is no possibility of misselection between tools. Its purpose (convert a raster image to SVG) is stated unambiguously.
The single tool follows a clear verb_noun convention (vectorize_image). With no other tools to conflict with, the naming pattern is fully consistent.
A one-tool surface is defensible for a narrowly focused image converter, but it is thin: the description itself points users to an external website for gradient fitting, posterize mode, and layer editing. The server feels like a minimal wrapper rather than a self-contained toolset.
The core operation (raster to SVG with color/bw modes) is covered, but there are notable gaps: no control over output path, tracing parameters (precision, color count), batch conversion, or the advanced features mentioned in the description. Agents can accomplish the basic workflow but hit dead ends for anything beyond it.
Create professional SVG artwork, icons and vector logos from a prompt, or vectorize any image.
Convert PNG, JPG and WebP to SVG. Recolor an SVG, reduce its colors or remove its background.
Generate images from Claude, ChatGPT, Cursor and Codex with 12 top AI image models and 10 tools: up to 4 images per call, SVG vectorization, editing and more. Remote MCP with OAuth, no API keys. Pay per image, no subscription.
Generate images from Claude, ChatGPT, Cursor and Codex with 12 top AI image models. Up to 4 images per call, SVG vectorization, background removal, upscaling and image-to-video. Remote MCP with OAuth, no API keys.
