Spronta MCP Server
Enables AI assistants to upload, transform, and serve images using the Spronta Image CDN. It supports real-time image processing, global CDN delivery, and management of projects and transform presets through natural language.
README
@spronta/mcp
MCP server for uploading, transforming, and serving images on a global CDN — directly from AI assistants.
Connect Claude, GPT, Cursor, Windsurf, or any MCP-compatible client to the Spronta Image CDN. Upload images from URLs or base64, apply real-time transforms (resize, crop, format conversion, smart crop, blurhash), create reusable presets, generate signed URLs, and monitor usage — all through natural language.
Why use this?
- Upload images from AI — give your LLM an image URL and it uploads, optimizes, and returns a CDN link
- Real-time transforms — resize, crop, convert to WebP/AVIF/JXL, smart crop with face detection
- Global CDN delivery — images served from edge locations worldwide
- Blurhash generation — automatic blur placeholders computed on upload
- Signed URLs — HMAC-SHA256 signed URLs with expiration for private images
- Named presets — create reusable transform configurations (e.g.
thumbnail,hero,og-image) - No vendor lock-in — MIT licensed, open source
Quick start
Install
npm install -g @spronta/mcp
# or use directly
npx @spronta/mcp
Claude Code
claude mcp add spronta \
-e SPRONTA_API_KEY=spronta_img_... \
-e SPRONTA_PROJECT_ID=your-project-id \
-- npx @spronta/mcp
Claude Desktop / Cursor / Windsurf
Add to your MCP configuration file:
{
"mcpServers": {
"spronta": {
"command": "npx",
"args": ["@spronta/mcp"],
"env": {
"SPRONTA_API_KEY": "spronta_img_...",
"SPRONTA_PROJECT_ID": "your-project-id"
}
}
}
}
Get your API key
- Sign up at app.spronta.com (free tier: 100 images)
- Create a project in the dashboard
- Copy your API key from project settings
Environment variables
| Variable | Required | Description |
|---|---|---|
SPRONTA_API_KEY |
Yes | Your project API key |
SPRONTA_PROJECT_ID |
No | Default project ID (can also be passed per-tool) |
SPRONTA_API_URL |
No | API base URL (default: https://app.spronta.com/api) |
18 tools available
Projects
| Tool | Description |
|---|---|
list_projects |
List all image projects |
create_project |
Create a new project |
get_project |
Get project details with usage stats |
update_project |
Update name or custom domain |
delete_project |
Permanently delete a project |
Upload & Images
| Tool | Description |
|---|---|
upload_image |
Upload from URL or base64 — handles presigned upload, blurhash generation, and CDN URL creation in one call |
list_images |
List images with pagination |
update_image |
Update alt text and tags |
delete_image |
Delete an image from CDN and storage |
Transform Presets
| Tool | Description |
|---|---|
list_presets |
List named transform presets |
create_preset |
Create a preset (use in CDN URLs with ?t=name) |
update_preset |
Update a preset's name or transforms |
delete_preset |
Delete a preset |
URL Signing
| Tool | Description |
|---|---|
get_signing_config |
Get URL signing configuration |
update_signing |
Enable/disable HMAC-SHA256 signing |
generate_signed_url |
Generate a signed CDN URL with optional expiration |
Analytics & Utility
| Tool | Description |
|---|---|
get_usage |
Get daily metrics (requests, bandwidth, transforms) |
build_cdn_url |
Build a CDN URL with transform parameters (no API call) |
Example prompts
Once connected, just ask your AI:
> Upload this image to my project: https://example.com/photo.jpg
> Create a thumbnail preset: 200x200, cover crop, face detection, webp format
> How many requests did my project handle this week?
> Generate a signed URL for hero.jpg that expires in 1 hour
> List all my images and update the alt text on the product shots
> Build a CDN URL for banner.png at 1200px wide in avif format
Transform options
When creating presets or building URLs, these transforms are available:
| Parameter | Type | Description |
|---|---|---|
width |
integer | Output width (1–8192) |
height |
integer | Output height (1–8192) |
fit |
string | cover, contain, fill, scale-down, crop, pad |
format |
string | auto, webp, avif, jpeg, png, jxl |
quality |
integer | 1–100 |
qualityMode |
string | high, medium, low (perceptual targeting) |
gravity |
string | auto, face, center, position keywords |
blur |
integer | Gaussian blur (1–250) |
sharpen |
number | Unsharp mask (0–10) |
radius |
integer | Corner radius (1–9999, or "max" for circle) |
grayscale |
boolean | Convert to grayscale |
brightness |
integer | -100 to 100 |
contrast |
number | -100 to 100 |
saturation |
integer | -100 to 100 |
sepia |
boolean | Sepia tone filter |
rotate |
integer | 90, 180, 270 |
flip |
string | h (horizontal), v (vertical) |
How upload works
The upload_image tool handles the full upload pipeline in a single call:
- You provide an image URL (or base64 data) and a filename
- The server fetches the image
- Requests a presigned upload URL from Spronta
- Uploads the file directly to storage
- Confirms the upload — Spronta computes the blurhash and detects dimensions
- Returns the confirmed image with CDN URL, blurhash, and metadata
Related packages
@spronta/images— Core URL builder SDK (zero deps)@spronta/images-react— React components@spronta/images-next— Next.js integration- WordPress plugin — Automatic CDN for WordPress
Links
License
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。
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。
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。
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。
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。
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。
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。