This guide provides step-by-step instructions for integrating the Altinity MCP Server with various AI tools and platforms.
- Install Helm Chart with Minikube and Helm from GHCR.io (Token Enabled)
- Generate Token from JWE Token Generator
- Integrate with Claude.ai Chat Web Interface
- Integrate with OpenAI ChatGPT GPTs
- Integrate with Claude Desktop
- Integrate with Claude Code
- Integrate with Cursor.com
- Integrate with Windsurf
To install the Altinity MCP Helm chart with token authentication enabled using Minikube:
-
First, ensure you have Minikube and Helm installed:
minikube start helm version
-
Install the Altinity MCP Helm chart from GHCR.io with token authentication enabled:
helm install altinity-mcp oci://ghcr.io/altinity/altinity-mcp/helm/altinity-mcp \ --set config.server.jwe.enabled=true \ --set config.server.jwe.jweSecretKey="your-jwe-secret-key" \ --set config.server.jwe.jwtSecretKey="your-jwt-secret-key" \ --set config.server.transport=http \ --set config.server.port=8080
-
Forward the port to access the service:
kubectl port-forward service/altinity-mcp 8080:8080
For detailed configuration options, see the Helm Chart README.
To generate a JWE token for ClickHouse authentication with 24h expiry using Docker:
docker run --rm ghcr.io/altinity/altinity-mcp:latest jwe-token-generator \
--jwe-secret-key "your-jwe-secret-key" \
--jwt-secret-key "your-jwt-secret-key" \
--host "clickhouse.example.com" \
--port 8123 \
--database "analytics" \
--username "user123" \
--password "password123" \
--expiry 86400This will output a JWE token that can be used for authentication with the Altinity MCP server.
To integrate Altinity MCP with Claude.ai:
- Navigate to https://claude.ai and log in to your account.
- Go to Settings > Connectors.
- In the Connectors settings, click on Add custom connector.
- Enter a name for your connector (e.g.,
altinity-mcp-jwe) and the server URL, which should include the JWE token. For example:https://host/token/http.
- After adding the connector, you can enable it for your chat session by clicking the tool icon.
- You can see the list of available tools and enable or disable them.
- Once enabled, you can use the tools in your chat. For example, you can ask Claude to list tables from your ClickHouse database.
- Claude will show the results from the tool execution.
- You can continue the conversation with Claude to analyze the data.
To integrate Altinity MCP with OpenAI GPTs:
- Navigate to https://chatgpt.com and log in to your account.
- Go to the GPTs section and click Create.
- In the configuration interface, go to the Configure tab.
- Scroll down to the Actions section and click Create new action.
- Import the OpenAPI schema from your Altinity MCP server URL. For example:
https://host/openapi.
- Configure the authentication with your JWE token using the API Key option.
- Test the connection in the preview mode to ensure it's working properly.
- You can now interact with your ClickHouse database using the GPT.
To integrate Altinity MCP with Claude Desktop:
- Download and Open the Claude Desktop application https://claude.ai/download.
- Go to Settings > Developer.
- In the Local MCP servers section, click on Edit Config.
- Add your Altinity MCP server details to the configuration file.
- Save the configuration with following format:
{
"mcpServers": {
"altinity-mcp": {
"command": "npx",
"args": ["-y", "mcp-remote", "http{s}://your-mcp-host:port/generated_jwe_token/http"]
}
}
}Download and install Claude Code https://www.anthropic.com/claude-code
To integrate Altinity MCP with Claude Code, you can use the claude mcp add command for HTTP transport:
claude mcp add --transport http altinity-mcp https://your-mcp-host:port/generated_jwe_token/httpReplace https://your-mcp-host:port/generated_jwe_token/http with the actual URL of your Altinity MCP server, including the JWE token and /http suffix.
This command will configure Claude Code to use your Altinity MCP server as a tool provider.
Go to: Settings -> Cursor Settings -> MCP -> Add new global MCP server
Pasting the following configuration into your Cursor ~/.cursor/mcp.json file is the recommended approach. You may also install in a specific project by creating .cursor/mcp.json in your project folder.
{
"mcpServers": {
"altinity-mcp": {
"url": "https://your-mcp-server-url:port/default/http",
"headers": {
"Authorization": "Bearer your-jwe-token"
}
}
}
}See Cursor MCP docs for more info.
To integrate Altinity MCP with Windsurf, you need to add a new entry to the mcpServers section in the IDE's settings. This will allow you to use the Altinity MCP server as a tool provider.
Add this to your Windsurf MCP config file. See Windsurf MCP docs for more info.
{
"mcpServers": {
"context7": {
"serverUrl": "https://your-mcp-server-url:port/default/http",
"headers": {
"Authorization": "Bearer your-jwe-token"
}
}
}
}If you encounter issues during integration:
- Verify your Altinity MCP server is running and accessible
- Ensure your JWE token is valid and not (use
curl -vvv https://your-mcp-host/your_jwe_token/http) - Check that the server configuration matches the integration settings
- Confirm network connectivity between the AI tool and your Altinity MCP server
- Review server logs for authentication or connection errors
For more detailed troubleshooting, refer to the main README.md file.
















