BPT MCP Server
A self-contained Model Context Protocol server that turns any MCP-aware AI client into a BlazorPowerTools expert.
A BPT expert, packaged as a Docker image
The BPT MCP Server bakes the entire BlazorPowerTools skill corpus into a lightweight Linux image (~50 MB) and exposes it over the Model Context Protocol's Streamable-HTTP transport. Connect Claude Desktop, Cline, Cursor, Continue, or any other MCP client and your AI assistant becomes a BPT specialist with grounded knowledge of every component, parameter, and pattern.
Download MCP Server v10.9.39 (53.2 MB)What's inside
Each release ships an MCP server with the complete BPT skill corpus baked in — both component-reference skills and the whole-app scaffolding skills:
Component reference
Every form control (CheckBox, Dropdown, Slider, …) plus BptRootComponent and the cascading error system.
The full 15-chart family (Bar, Line, Pie, Donut, Scatter, Bubble, Radar, Polar, TreeMap, Funnel, Histogram, BoxWhisker, Bullet, DateAxis, Trendline) with tooltips, palettes, 3D & responsive sizing.
Image, AnimationEffect, Background, Carousel, ObjectDetector.
ServerMonitor, TerminalClient, ThemeBuilder, LandingPage, ImageEditor, and the BptAppFramework responsive shell.
Render diagnostics, browser size, element tracking.
Speech, transcription, text generation, the reusable BptAiChatPanel.
The full BptImageEditor reference — 25+ tools, layer system, 85+ JS API functions.
The BptLandingPage visual builder — Edit/Preview modes, multi-page projects, element catalog.
Whole-app scaffolding — the design system
The "design my app with BPT" wizard — 9-question interview, implementation plan, dry-run preview. See the walkthrough →
The 10-section runbook the wizard follows — written to .claude/plans/bpt-scaffold.md for review before any file is created.
How to run
- Download the zip above and extract it.
- Load the image into your local Docker engine:
cd bpt-mcp-server-10.9.39 bash load.sh - Start the server:
docker run --rm -p 7180:7180 bpt-mcp-server:10.9.39 - Verify it's healthy:
curl http://localhost:7180/healthshould return{"status":"ok","skills":11}(10 skill files plus the index). - Connect any MCP client to
http://localhost:7180/mcp(see the snippets below).
MCP client configuration
Claude Desktop / Cline / Cursor / Continue
Add an entry to your client's MCP config:
{
"mcpServers": {
"bpt-expert": {
"url": "http://localhost:7180/mcp",
"transport": "http"
}
}
}
Once connected, your assistant has full access to:
- Every BPT skill as an MCP resource (URI scheme
bpt-skill://) - The
bpt_expert_guidanceprompt, which injects every skill as grounding context for any BPT question
Smoke test with the official MCP Inspector
npx @modelcontextprotocol/inspector
Connect the inspector to http://localhost:7180/mcp. The Resources tab should list every skill file; the Prompts tab should show bpt_expert_guidance.
Security
By default, the server binds to 0.0.0.0 with no auth — fine for a developer running
it locally on a single workstation. If the container is reachable beyond localhost:
- Bind to localhost only:
-p 127.0.0.1:7180:7180 - Or require a bearer token:
-e MCP_AUTH_TOKEN="$(openssl rand -hex 24)"and add"headers": { "Authorization": "Bearer YOUR_TOKEN" }to your client config.
New here? See it in action first
Developing with the BPT MCP Server walks through the “design my app with BPT” wizard end to end — the 9 questions, a real worked example with its answers, the exact file tree it plans, and the patterns & best practices you get. See precisely what lands in your project before you connect a single client.