sisense
MCP server enabling LLMs to query Sisense data sources and build charts via natural language prompts.
claude mcp add --transport stdio sisense-sisense-mcp-server node path/to/server.js \ --env PORT="Optional server port, defaults to 3001" \ --env SISENSE_URL="Full URL to your Sisense instance (e.g., https://instance.sisense.com)" \ --env SISENSE_TOKEN="Sisense API authentication token"
How to use
This Sisense MCP Server provides a lightweight, per-session authenticated bridge for LLMs to interact with Sisense data models. It exposes three MCP tools: getDataSources, which retrieves available Sisense data sources or data models; getDataSourceFields, which lists all fields for a specified data source; and buildChart, which converts natural language prompts into chart configurations that render in an MCP Apps iframe when supported by the client. The server supports streaming via HTTP/SSE for compatible clients and derives its public URL from request headers, making it suitable to run behind proxies or tunneling services like ngrok. You can connect using Sisense credentials supplied via URL parameters or environment variables (SISENSE_URL and SISENSE_TOKEN).
How to install
Prerequisites:
- Bun >= 1.0.0 (recommended) or Node.js >= 18.0.0
- Sisense instance with API access
- Sisense API token
Installation steps:
- Ensure Bun (or Node) is installed on your system.
- Bun: https://bun.sh/
- Node: https://nodejs.org/
- Clone the repository or download the source.
- Install dependencies:
- bun install (If using Node, run: npm install or yarn install after ensuring the project has a package.json)
- Start the server in development mode:
- bun run dev
- If using Node/npm scripts, ensure a start script is defined, e.g., node dist/server.js or node src/server.js after build.
- Verify the server is running. The server will typically listen on port 3001 by default and expose health and endpoints as described in the README.
Additional notes
- The server uses per-session authentication via Sisense credentials passed through URL parameters (sisenseUrl and sisenseToken) or via environment variables (SISENSE_URL and SISENSE_TOKEN). If both methods are provided, URL parameters take precedence.
- Recommended environment handling: Do not commit credentials to version control. Use secret managers or vaults in production.
- Do not bind to 0.0.0.0 in production; prefer localhost or a reverse proxy/socket.
- When running behind proxies, the server derives its public base URL from request headers.
- Endpoints of interest include health at /health and visuals at /screenshots/ if supported by the build.
- The three MCP Tools (getDataSources, getDataSourceFields, buildChart) are designed to integrate with Claude and other MCP Apps-enabled clients where buildChart can render an interactive View inside an iframe.
Related MCP Servers
obsidian -tools
Add Obsidian integrations like semantic search and custom Templater prompts to Claude or any MCP client.
mcp
Octopus Deploy Official MCP Server
furi
CLI & API for MCP management
mcp -arangodb
This is a TypeScript-based MCP server that provides database interaction capabilities through ArangoDB. It implements core database operations and allows seamless integration with ArangoDB through MCP tools. You can use it wih Claude app and also extension for VSCode that works with mcp like Cline!
CodeRAG
Advanced graph-based code analysis for AI-assisted software development
mcp-bundler
Is the MCP configuration too complicated? You can easily share your own simplified setup!