icogenie-mcp
AI-powered SVG icon generation MCP server. Generate production-ready SVG icons from text descriptions with customizable styles, sizes, and themes.
README
@icogenie/mcp
MCP (Model Context Protocol) server for IcoGenie. Enables AI agents like Claude to generate production-ready SVG icons programmatically.
Installation
npm install -g @icogenie/mcp
# or use directly with npx
npx @icogenie/mcp
Configuration
Claude Desktop
Add to your Claude Desktop configuration (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"icogenie": {
"command": "npx",
"args": ["-y", "@icogenie/mcp"]
}
}
}
Claude Code
The server works automatically when installed globally or via npx.
Authentication
On first use, the MCP server will:
- Open your browser for authentication
- Ask you to approve access
- Save the session token to
~/.icogenie/config.json
Subsequent uses are automatic - the token is shared with the IcoGenie CLI.
CI/CD Environments
Set ICOGENIE_SESSION_TOKEN environment variable to skip browser authentication:
export ICOGENIE_SESSION_TOKEN="your-session-token"
Available Tools
generate_icon
Generate a single icon preview from a text description.
Cost: 1 credit
generate_icon({
prompt: "home icon",
style: "solid", // or "outline"
variations: 1, // 1, 2, or 4
referenceImagePath: "/path/to/reference.png" // optional
})
Returns: { sessionId, preview, previews, creditsRemaining, sessionData, suggestions }
regenerate_icon
Regenerate a specific icon variation. Returns 4 candidate previews — use confirm_regeneration to finalize your choice (two-phase flow).
Cost: 1 credit
regenerate_icon({
sessionId: "abc123", // for single icons
bundleId: "xyz789", // for bundles (use one or the other)
index: 0, // which variation (0-based)
prompt: "Make it more 3D" // optional refinement
})
Returns: { candidates, regenToken, creditsRemaining }
confirm_regeneration
Finalize a regeneration by selecting one of the 4 candidates.
Cost: Free (included in regenerate_icon)
confirm_regeneration({
regenToken: "tok_abc123",
selectedIndex: 2 // 0-3, which candidate to keep
})
Returns: { success, preview, creditsRemaining }
check_credits
Check your current credit balance.
Cost: Free
check_credits()
Returns: { credits, team, user }
download_icon
Download the final SVG + PNG package for an icon or bundle.
Cost: 5 credits (single) or 4 credits/icon (bundle)
download_icon({
generationId: "abc123", // or bundleId
outputPath: "./icons.zip" // optional, returns base64 if omitted
})
normalize_bundle
Plan an icon bundle by generating an AI-enhanced icon list from a description.
Cost: Free (rate-limited)
normalize_bundle({
description: "food delivery app",
targetCount: 10,
style: "solid"
})
Returns: { bundleId, icons, bundleType, styleRecommendation, reasoning }
generate_bundle
Generate a bundle of icons from an icon list.
Cost: 1 credit per icon
generate_bundle({
icons: [
{ name: "home", description: "Home navigation icon" },
{ name: "cart", description: "Shopping cart icon" }
],
style: "solid",
bundleId: "bundle_abc123" // optional, from normalize_bundle
})
Returns: { bundleId, iconCount, icons: [{ name, description, preview }], pricing, creditsUsed, creditsRemaining }
save_to_library
Save a generated icon to your personal library.
Cost: Free
save_to_library({
generationId: "abc123",
name: "notification-bell", // optional display name
tags: ["ui", "alerts"] // optional tags
})
Returns: { success, libraryItemId }
list_library
List icons saved in your library.
Cost: Free
list_library({
page: 1, // optional, default 1
limit: 20, // optional, default 20
tag: "ui" // optional filter
})
Returns: { items: [{ id, name, tags, preview, createdAt }], total, page }
download_from_library
Download an icon from your library as SVG + PNG.
Cost: Free
download_from_library({
libraryItemId: "lib_abc123",
outputPath: "./bell-icon.zip" // optional, returns base64 if omitted
})
Returns: { savedTo } or { base64, filename }
claim_daily_credits
Claim 2 free credits, available once every 24 hours.
Cost: Free
claim_daily_credits()
Returns: { credited, creditsRemaining, nextClaimAt }
Example Workflow
-
Check credits:
check_credits() → { credits: 50, team: { name: "Personal" } } -
Generate a single icon:
generate_icon({ prompt: "notification bell icon", style: "outline" }) → { sessionId: "abc123", preview: "...", creditsRemaining: 49 } -
Refine if needed (two-phase):
regenerate_icon({ sessionId: "abc123", index: 0, prompt: "Add a dot indicator" }) → { candidates: ["...", "...", "...", "..."], regenToken: "tok_xyz", creditsRemaining: 48 } confirm_regeneration({ regenToken: "tok_xyz", selectedIndex: 1 }) → { success: true, preview: "...", creditsRemaining: 48 } -
Download final package:
download_icon({ generationId: "abc123", outputPath: "./bell-icon.zip" }) → { savedTo: "./bell-icon.zip" }
Bundle Workflow
-
Plan the bundle (free):
normalize_bundle({ description: "e-commerce app icons", targetCount: 8 }) → { icons: [{ name: "cart", ... }, ...], styleRecommendation: "outline" } -
Review and generate (pass bundleId for traceability):
generate_bundle({ icons: [...], style: "outline", bundleId: "bundle_abc" }) → { bundleId: "bundle_abc", icons: [...], credits: { previewUsed: 8 } } -
Download bundle:
download_icon({ bundleId: "xyz789", outputPath: "./ecommerce-icons.zip" })
Environment Variables
| Variable | Description | Default |
|---|---|---|
ICOGENIE_API_URL |
API endpoint | https://www.icogenie.xyz |
ICOGENIE_SESSION_TOKEN |
Session token (for CI/CD) | - |
ICOGENIE_CONFIG_DIR |
Config directory | ~/.icogenie |
Credit Pricing
| Action | Cost |
|---|---|
| Preview (single) | 1 credit |
| Preview (bundle) | 1 credit/icon |
| Download (single) | 5 credits |
| Download (bundle) | 4 credits/icon |
| Regenerate | 1 credit |
| Daily claim | 2 free credits/day |
| Library ops | Free |
Purchase credits at www.icogenie.xyz.
When to Use MCP vs CLI vs Web
| Use Case | Best Option |
|---|---|
| Generate icons inside Claude/Cursor | MCP |
| Automate in CI/CD pipelines | CLI |
| Quick one-off generation | Web |
| Batch generate icon bundles | CLI or MCP |
| Browse and manage icon library | Web |
Related
- @icogenie/cli - Command-line interface
- IcoGenie Web - Web application
- AI Agent Docs - Machine-readable documentation
- Developer Docs - Integration guides
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。