@vemdev/mcp-server
Gives AI agents durable project memory via the Model Context Protocol, allowing them to read tasks, record decisions, search context, and sync snapshots to the cloud.
README
@vemdev/mcp-server
Official VEM MCP Server — gives AI agents (Claude, Cursor, Copilot, etc.) durable project memory via the Model Context Protocol.
Agents can read tasks, record decisions, search context, and sync snapshots to the cloud — all through structured MCP tools.
Quick start
npx @vemdev/mcp-server@latest --api-key <your-api-key>
Get your API key from app.vem.dev → Settings → API Keys.
MCP host configuration
Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"vem": {
"command": "npx",
"args": ["@vemdev/mcp-server@latest", "--api-key", "nk_your_key_here"]
}
}
}
Cursor / VS Code
Add to your MCP settings:
{
"vem": {
"command": "npx",
"args": ["@vemdev/mcp-server@latest", "--api-key", "nk_your_key_here"]
}
}
Authentication
There are three ways to provide your API key, in priority order:
1. --api-key flag (recommended for MCP hosts)
Pass the key directly when starting the server:
npx @vemdev/mcp-server@latest --api-key nk_your_key_here
The key is saved to ~/.vem/config.json on first run, so subsequent starts
without the flag will reuse it automatically.
2. VEM_API_KEY environment variable
Useful in CI or when you prefer not to store the key on disk:
VEM_API_KEY=nk_your_key_here npx @vemdev/mcp-server@latest
Or in your MCP host config:
{
"mcpServers": {
"vem": {
"command": "npx",
"args": ["@vemdev/mcp-server@latest"],
"env": {
"VEM_API_KEY": "nk_your_key_here"
}
}
}
}
3. vem login CLI (if you already use the CLI)
If you have the VEM CLI installed and have already logged in, the MCP server picks up the saved key automatically — no extra config needed:
npx @vemdev/cli@latest login nk_your_key_here
npx @vemdev/mcp-server@latest # key is read from ~/.vem/config.json
Getting an API key
- Go to app.vem.dev and sign in
- Navigate to Settings → API Keys
- Click Create key and copy the
nk_...token - Pass it to the server via any of the methods above
Available tools
| Tool | Description |
|---|---|
get_active_tasks |
List all active tasks in the project |
add_task |
Create a new task |
start_task |
Transition a task to in-progress |
complete_task |
Mark a task as done with evidence |
update_task |
Update task fields (title, status, priority) |
delete_task |
Delete a task |
get_task_details |
Get full details of a specific task |
get_task_context |
Get ephemeral context for a task |
update_task_context |
Update ephemeral context for a task |
get_context |
Read CONTEXT.md (project summary) |
update_current_state |
Update CURRENT_STATE.md |
list_decisions |
List recorded architectural decisions |
add_decision |
Record a new architectural decision |
search_memory |
Semantic search across project memory |
ask_question |
Ask a question about the project context |
get_changelog |
Read recent changelog entries |
apply_vem_update |
Apply a structured vem_update block |
sync_push |
Push a memory snapshot to the cloud |
sync_pull |
Pull the latest snapshot from the cloud |
get_subtasks |
List subtasks of a task |
list_agent_sessions |
List recorded agent sessions |
save_session_stats |
Save session stats for a task |
Environment variables
| Variable | Default | Description |
|---|---|---|
VEM_API_KEY |
— | API key (alternative to --api-key flag) |
VEM_API_URL |
https://api.vem.dev |
VEM API endpoint |
VEM_AGENT_NAME |
mcp-agent |
Agent name shown in the dashboard |
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。
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。