
Mcp Icon Visual
Connects Claude to Iconify's icon library with visual output designed for multimodal models. Exposes search_icons to query by keyword and get back a 4x4 PNG grid where Claude can see the results and reference them by cell number, plus get_svg to retrieve raw SVG markup from Iconify IDs, URLs, or local files. Also includes render_grid for side-by-side comparisons, useful when you're iterating on SVG edits or checking style consistency across a set. Built in Rust with stdio transport. The key trick is returning images instead of metadata so vision models can actually judge icon style and fit rather than guessing from filenames.
- GitHub stars
- 1
- tools
- 3
- transport
- STDIO
- last updated
- Mar 11, 2026
Tools
3 toolsPublic tool metadata for what this MCP can expose to an agent.
search_iconsSemantic icon search. Returns a 4x4 PNG grid (512x512, cells 0-15) and a mapping of cell numbers to SVG URLs.
Semantic icon search. Returns a 4x4 PNG grid (512x512, cells 0-15) and a mapping of cell numbers to SVG URLs.
No parameter schema in public metadata yet.
get_svgRetrieve raw SVG markup from any source. Returns the complete SVG string for inspection or modification.
Retrieve raw SVG markup from any source. Returns the complete SVG string for inspection or modification.
No parameter schema in public metadata yet.
render_gridRender 1-16 SVGs onto a 4x4 comparison grid (512x512, cells 0-15). Useful for comparing icon variants side-by-side or previewing local icons alongside search results.
Render 1-16 SVGs onto a 4x4 comparison grid (512x512, cells 0-15). Useful for comparing icon variants side-by-side or previewing local icons alongside search results.
No parameter schema in public metadata yet.
Sponsors
Readme
mcp-icon-visual
MCP server that gives AI agents the ability to search, retrieve, compare, and render SVG icons. Built with Rust and the rmcp SDK.
Combines local filesystem access with a remote rendering API to resolve SVGs from any source — inline strings, URLs, local files, or Iconify IDs — and render them onto visual comparison grids.
The key insight: every tool returns images. A vision-capable LLM doesn't just get metadata — it sees the icons. This turns icon selection from guesswork into a visual conversation.
Details
- Repository
- github.com/n0isy/brain-mcp-icon-visual
- Categories
- Design & Creative, Search & Web Crawling
- Package
@br-ai-n/mcp-icon-visual- Transport
- STDIO
- Registry
- MCP registry, active