Deerdawn MCP server
{
"mcpServers": {
"deerdawn": {
"command": "npx",
"args": [
"-y",
"@deerdawn/mcp-server@latest"
],
"env": {
"DEERDAWN_API_KEY": "<deerdawn_api_key>",
"DEERDAWN_API_URL": "https://...",
"DEERDAWN_SURFACE_ID": "<deerdawn_surface_id>"
}
}
}
}{
"mcpServers": {
"deerdawn": {
"command": "npx",
"args": [
"-y",
"@deerdawn/mcp-server@latest"
],
"env": {
"DEERDAWN_API_KEY": "<deerdawn_api_key>",
"DEERDAWN_API_URL": "https://...",
"DEERDAWN_SURFACE_ID": "<deerdawn_surface_id>"
}
}
}
}{
"mcpServers": {
"deerdawn": {
"command": "npx",
"args": [
"-y",
"@deerdawn/mcp-server@latest"
],
"env": {
"DEERDAWN_API_KEY": "<deerdawn_api_key>",
"DEERDAWN_API_URL": "https://...",
"DEERDAWN_SURFACE_ID": "<deerdawn_surface_id>"
}
}
}
}{
"servers": {
"deerdawn": {
"command": "npx",
"args": [
"-y",
"@deerdawn/mcp-server@latest"
],
"env": {
"DEERDAWN_API_KEY": "${input:deerdawn_api_key}",
"DEERDAWN_API_URL": "${input:deerdawn_api_url}",
"DEERDAWN_SURFACE_ID": "${input:deerdawn_surface_id}"
}
}
},
"inputs": [
{
"id": "deerdawn_api_key",
"type": "promptString",
"description": "Deerdawn api key",
"password": true
},
{
"id": "deerdawn_api_url",
"type": "promptString",
"description": "Deerdawn api url",
"password": true
},
{
"id": "deerdawn_surface_id",
"type": "promptString",
"description": "Deerdawn surface id",
"password": true
}
]
}{
"mcpServers": {
"deerdawn": {
"command": "npx",
"args": [
"-y",
"@deerdawn/mcp-server@latest"
],
"env": {
"DEERDAWN_API_KEY": "<deerdawn_api_key>",
"DEERDAWN_API_URL": "https://...",
"DEERDAWN_SURFACE_ID": "<deerdawn_surface_id>"
}
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall npx -y @deerdawn/mcp-serverversion 1.0.53
What it does
DeerDawn is an MCP server that provides AI session memory, maintaining context about your project across sessions so new AI conversations start with relevant background information already loaded. It works with Claude Code, Claude Desktop, ChatGPT, Cursor, Windsurf, VS Code, Zed, and other MCP-compatible tools.
- Start a new AI session in any tool and immediately have the AI briefed on what you were working on, decisions made, and current blockers.
- Automatically capture project context and flush it when sessions end, reducing repetition across multiple conversations.
- Write session context to `.deerdawn-context.md` or `.cursorrules` files in your repo so context loads without explicit tool calls.
Tools this server exposes
| Tool | Does |
|---|---|
| start_session | Begin a new AI session with your project brief already loaded. |
| get_auth_status | Check authentication status and retrieve the sign-in URL if needed. |
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.