CallRemind MCP Server
Enables AI agents to make instant calls, manage CEMAS emergency notifications, pick Edge TTS voices, manage agents, leads, scheduled calls, and query call history via the CallRemind REST API.
README
CallRemind MCP Server
MCP (Model Context Protocol) server that lets AI agents interact with the CallRemind REST API — place instant calls, manage CEMAS emergency notifications, pick Edge TTS voices, manage agents/leads/scheduled calls, and query call history.
Targets the CallRemind REST API (https://api.callremind.my/v1), not the
outbound engine.
Table of contents
Requirements
- Node.js >= 18
- A CallRemind API key (your user
x-api-key). Get it from your CallRemind profile, or the admin/v1/admin/generate-keyendpoint.
Two ways to run
A. npm package (easiest)
Install the CLI (stdio) and run it directly:
npm install -g callremind-mcp
CALLREMIND_API_KEY=your_key callremind-mcp
For an HTTP/SSE server (remote clients like ChatGPT):
CALLREMIND_API_KEY=your_key CALLREMIND_MCP_PORT=8921 npx callremind-mcp index-sse.js
B. From source / Docker
git clone git@github.com:callremind/cemas-mcp.git
cd cemas-mcp
npm install
cp .env.example .env # fill in CALLREMIND_API_URL + CALLREMIND_API_KEY
npm start # stdio (local clients)
npm run start:sse # HTTP/SSE on http://0.0.0.0:8921/mcp
Docker:
docker build -t callremind-mcp .
# stdio
docker run --rm -e CALLREMIND_API_KEY=your_key callremind-mcp
# HTTP/SSE
docker run --rm -p 8921:8921 -e CALLREMIND_API_KEY=your_key callremind-mcp node index-sse.js
Configuration
| Env var | Default | Description |
|---|---|---|
CALLREMIND_API_URL |
https://api.callremind.my/v1 |
CallRemind REST API base |
CALLREMIND_API_KEY |
— | Your x-api-key (required) |
CALLREMIND_MCP_PORT |
8921 |
HTTP/SSE port (index-sse.js only) |
CALLREMIND_MCP_HOST |
0.0.0.0 |
HTTP/SSE bind host |
CALLREMIND_MCP_ENDPOINT |
/mcp |
HTTP/SSE MCP endpoint path |
Client setup
Claude Desktop
Edit claude_desktop_config.json (Claude Desktop → Settings → Developer →
Edit Config):
{
"mcpServers": {
"callremind": {
"command": "npx",
"args": ["-y", "callremind-mcp"],
"env": { "CALLREMIND_API_KEY": "your_key" }
}
}
}
Claude Code
claude mcp add callremind -- npx -y callremind-mcp
# or with env:
claude mcp add callremind -e CALLREMIND_API_KEY=your_key -- npx -y callremind-mcp
Codex (OpenAI CLI)
codex mcp add callremind -- npx -y callremind-mcp
# pass the key via shell env so it reaches the child process:
CALLREMIND_API_KEY=your_key codex
ChatGPT (remote, HTTP/SSE)
Run the SSE server somewhere reachable, then point ChatGPT at it:
CALLREMIND_API_KEY=your_key CALLREMIND_MCP_PORT=8921 node index-sse.js
In ChatGPT, add a custom MCP server with URL:
http://<your-host>:8921/mcp
The server speaks the Streamable HTTP transport (the modern MCP HTTP/SSE protocol), so any MCP-over-HTTP client can connect.
opencode
opencode.json (or ~/.config/opencode/opencode.json):
{
"mcp": {
"callremind": {
"type": "local",
"command": ["npx", "-y", "callremind-mcp"],
"environment": { "CALLREMIND_API_KEY": "your_key" },
"enabled": true
}
}
}
Cursor
Cursor → Settings → MCP → Add server:
{
"mcpServers": {
"callremind": {
"command": "npx",
"args": ["-y", "callremind-mcp"],
"env": { "CALLREMIND_API_KEY": "your_key" }
}
}
}
Generic HTTP/SSE clients
For any MCP client that supports remote HTTP servers:
Transport: Streamable HTTP
URL: http://<host>:8921/mcp
Tools
| Tool | Endpoint | Purpose |
|---|---|---|
makeInstantCall |
POST /v1/calls/instant |
Immediate outbound AI call |
getCemasConfig |
GET /v1/cemas |
CEMAS config + voice prefs |
saveCemasConfig |
PUT /v1/cemas/destination |
Save destination + voice/prosody |
listEdgeVoices |
GET /v1/tts/edge-voices |
List Edge TTS voices |
previewEdgeVoice |
POST /v1/tts/edge-preview |
Synthesize voice preview |
listVoiceClones / createVoiceClone |
/v1/voice-clones |
Manage voice clones |
listAgents / createAgent / updateAgent |
/v1/agents |
Manage AI agents |
listLeads / createLead |
/v1/contacts |
Manage contacts/leads |
listScheduledCalls / createScheduledCall |
/v1/scheduled-calls |
Schedule calls |
listCallLogs |
/v1/call-logs |
Call history |
getReports |
/v1/reports |
Call analytics |
getProfile / updateProfile |
/v1/profile |
Profile management |
listWebhookConfigurations / upsertWebhookConfiguration |
/v1/webhooks/... |
Webhooks |
Testing
npm test # or: node test-api.js (hits /v1/profile with your key)
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。