Pcd MCP server
{
"mcpServers": {
"pcd": {
"type": "http",
"url": "https://mcp.portablecontext.ai/mcp"
}
}
}{
"mcpServers": {
"pcd": {
"type": "http",
"url": "https://mcp.portablecontext.ai/mcp"
}
}
}{
"mcpServers": {
"pcd": {
"type": "http",
"url": "https://mcp.portablecontext.ai/mcp"
}
}
}{
"servers": {
"pcd": {
"type": "http",
"url": "https://mcp.portablecontext.ai/mcp"
}
}
}{
"mcpServers": {
"pcd": {
"type": "http",
"url": "https://mcp.portablecontext.ai/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 0.4.0
What it does
PCD (Portable Context Deck) is a hosted MCP server that lets users load structured personal context—knowledge, work state, skills, and signals—into any MCP-compatible AI client through a single endpoint. It is aimed at developers and knowledge workers who switch between multiple AI tools and want their context available consistently across all of them.
- Load your full working context into a new AI client session with a single command so it is already briefed on your current work.
- Maintain live work state and accumulated knowledge across different AI clients such as desktop apps, IDEs, and chat interfaces without re-entering context each time.
- Organize domain- or task-specific contexts as specialists and switch between them depending on what you are working on.
Tools this server exposes
| Tool | Does |
|---|---|
| load | Briefs the AI with the user's operating context or enters a specific container or specialist. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.