Integrations
Connect arifOS MCP to your preferred AI client. Select your integration below for step-by-step configuration.
Claude Desktop
Add arifOS MCP to Claude Desktop by editing your claude_desktop_config.json:
macOS
# Open the config file
code ~/Library/Application\ Support/Claude/claude_desktop_config.json
Windows
# Open the config file
code %APPDATA%\Claude\claude_desktop_config.json
Configuration
{
"mcpServers": {
"arifosmcp": {
"command": "python",
"args": ["-m", "arifosmcp"],
"env": {
"BRAVE_API_KEY": "your_brave_api_key_here"
}
}
}
}
Docker variant
{
"mcpServers": {
"arifosmcp": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-e", "BRAVE_API_KEY=your_key_here",
"ghcr.io/ariffazil/arifosmcp:latest"
]
}
}
}
Restart Claude Desktop after saving. The arifOS MCP tools will appear in the tools panel.
Cursor
Add arifOS MCP to Cursor via the MCP settings:
- Open Cursor Settings → Features → MCP Servers
- Click "Add new MCP Server"
- Use the following configuration:
{
"mcpServers": {
"arifosmcp": {
"command": "python",
"args": ["-m", "arifosmcp"],
"env": {
"BRAVE_API_KEY": "your_brave_api_key_here"
}
}
}
}
Alternatively, create a .cursor/mcp.json file in your project root:
{
"mcpServers": {
"arifosmcp": {
"command": "python",
"args": ["-m", "arifosmcp"],
"env": {
"BRAVE_API_KEY": "your_brave_api_key_here"
}
}
}
}
VS Code
For VS Code with Copilot Chat or other MCP-compatible extensions:
{
"servers": {
"arifosmcp": {
"type": "stdio",
"command": "python",
"args": ["-m", "arifosmcp"],
"env": {
"BRAVE_API_KEY": "your_brave_api_key_here"
}
}
}
}
Or add to your VS Code settings.json:
{
"mcp.servers": {
"arifosmcp": {
"type": "stdio",
"command": "python",
"args": ["-m", "arifosmcp"],
"env": {
"BRAVE_API_KEY": "your_brave_api_key_here"
}
}
}
}
ChatGPT
Connect arifOS MCP to ChatGPT using the HTTP transport and the ChatGPT Actions configuration:
- First, start the arifOS MCP server in HTTP mode:
arifosmcp serve --transport http --port 8080 --host 0.0.0.0
- In ChatGPT, go to Settings → Actions → Create new action
- Set the API endpoint to your server URL
- Import the OpenAPI schema from
/openapi.json
Kimi
Connect arifOS MCP to Kimi (Moonshot AI) using the MCP plugin system:
- Open Kimi's plugin settings
- Select "Add Custom MCP Server"
- Configure the STDIO connection:
{
"name": "arifosmcp",
"transport": "stdio",
"command": "python",
"args": ["-m", "arifosmcp"],
"env": {
"BRAVE_API_KEY": "your_brave_api_key_here"
}
}
Python SDK
Use the MCP Python SDK to connect programmatically:
pip install mcp arifosmcp
import asyncio
from mcp import ClientSession, StdioServerParameters
from mcp.client.stdio import stdio_client
async def main():
server = StdioServerParameters(
command="python",
args=["-m", "arifosmcp"],
env={"BRAVE_API_KEY": "your_key"}
)
async with stdio_client(server) as (read, write):
async with ClientSession(read, write) as session:
await session.initialize()
# List available tools
tools = await session.list_tools()
print(f"Available tools: {[t.name for t in tools.tools]}")
# Call reality_compass
result = await session.call_tool(
"reality_compass",
{"input": "quantum computing 2026", "mode": "search"}
)
print(result.content[0].text)
# Call arifOS_kernel
result = await session.call_tool(
"arifOS_kernel",
{
"query": "Explain quantum entanglement",
"risk_tier": "low",
"use_critique": True
}
)
print(result.content[0].text)
asyncio.run(main())
HTTP Client
from mcp import ClientSession
from mcp.client.streamable_http import streamablehttp_client
async def main():
async with streamablehttp_client("http://localhost:8080/mcp") as (read, write, _):
async with ClientSession(read, write) as session:
await session.initialize()
result = await session.call_tool("check_vital", {})
print(result.content[0].text)
asyncio.run(main())
TypeScript SDK
Use the MCP TypeScript SDK:
npm install @modelcontextprotocol/sdk
// TypeScript example using MCP SDK
import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StdioClientTransport } from "@modelcontextprotocol/sdk/client/stdio.js";
const transport = new StdioClientTransport({
command: "python",
args: ["-m", "arifosmcp"],
env: { BRAVE_API_KEY: "your_key" }
});
const client = new Client({
name: "my-app",
version: "1.0.0"
});
await client.connect(transport);
// List tools
const tools = await client.listTools();
console.log("Tools:", tools.tools.map(t => t.name));
// Call reality_compass
const result = await client.callTool({
name: "reality_compass",
arguments: {
input: "arifOS MCP documentation",
mode: "search"
}
});
console.log(result.content);
curl / HTTP
Interact with arifOS MCP directly via HTTP (requires HTTP transport mode):
Start the server
arifosmcp serve --transport http --port 8080
Check system health
curl -X POST http://localhost:8080/mcp \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "check_vital",
"arguments": {}
}
}'
Search with reality_compass
curl -X POST http://localhost:8080/mcp \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 2,
"method": "tools/call",
"params": {
"name": "reality_compass",
"arguments": {
"input": "quantum computing breakthroughs 2026",
"mode": "search",
"fetch_top_k": 2
}
}
}'
Run the governance kernel
curl -X POST http://localhost:8080/mcp \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 3,
"method": "tools/call",
"params": {
"name": "arifOS_kernel",
"arguments": {
"query": "Is this deployment safe?",
"risk_tier": "high",
"use_critique": true,
"requested_persona": "auditor"
}
}
}'
List available tools
curl -X POST http://localhost:8080/mcp \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 4,
"method": "tools/list",
"params": {}
}'