Alkemi MCP Server

1# Alkemi MCP Server
2 
3Integrate your Alkemi Data, connected to Snowflake, Google BigQuery, DataBricks and other sources, with your MCP Client.
4 
5This is a STDIO wrapper for the Streamable HTTP MCP Endpoint:
6 
7```
8https://api.alkemi.cloud/mcp
9```
10 
11Get your free API key at [datalab.alkemi.ai](https://datalab.alkemi.ai?ref=alkemi-mcp)
12 
13## Alkemi.ai
14 
15Querying databases requires knowledge about the schema of the tables and may require examples of the kinds of queries that can answer specific questions. Otherwise, you may be getting the wrong answers. Maintaining all that information in every agent or MCP Client that queries your database is a challenge and doesn't scale to teams looking to share data.
16 
17The Alkemi MCP Server uses Alkemi to store the database metadata, generate proper queries and actually query the database so you can share your MCP Server with teammates and everyone will have the same ability to query with quality.
18 
19 
20## Installation
21 
22To add OpenAI to Claude Desktop, add the server config:
23 
24On MacOS: `~/Library/Application Support/Claude/claude_desktop_config.json`
25 
26On Windows: `%APPDATA%/Claude/claude_desktop_config.json`
27 
28### Env Vars
29 
30- `MCP_NAME`: The name of the MCP Server. This is optional. If you configure multiple, this is required so they do not have the same names in your MCP Client..
31- `BEARER_TOKEN`: The Bearer token for the Streamable HTTP MCP Server. This is required for the STDIO MCP Integration.
32- `PRODUCT_ID`: The ID of the Product if you want to narrow scope to just a single product. This is optional.
33 
34 
35### Configuration
36 
37You can use it via `npx` in your Claude Desktop configuration like this:
38 
39```json
40{
41  "mcpServers": {
42    "alkemi": {
43      "command": "npx",
44      "args": [
45        "@alkemiai/alkemi-mcp"
46      ],
47      "env": {
48        "BEARER_TOKEN": "sk-12345"
49      }
50    }
51  }
52}
53```
54 
55 
56Or, if you clone the repo, you can build and use in your Claude Desktop configuration like this:
57 
58 
59```json
60 
61{
62  "mcpServers": {
63    "alkemi-data": {
64      "command": "node",
65      "args": [
66        "/path/to/alkemi-mcp/build/index.js"
67      ],
68      "env": {
69        "BEARER_TOKEN": "sk-12345"
70      }
71    }
72  }
73}
74```
75 
76If you want to specify a specific product that the MCP Server should use, you can specify the `PRODUCT_ID` environment variable. And with setting the `MCP_NAME`, you can configure multiple.
77 
78 
79```json
80 
81{
82  "mcpServers": {
83    "alkemi-customer-data": {
84      "command": "node",
85      "args": [
86        "/path/to/alkemi-mcp/build/index.js"
87      ],
88      "env": {
89        "MCP_NAME": "customer-data",
90        "PRODUCT_ID": "123",
91        "BEARER_TOKEN": "sk-12345"
92      }
93    },
94    "alkemi-web-traffic-data": {
95      "command": "node",
96      "args": [
97        "/path/to/alkemi-mcp/build/index.js"
98      ],
99      "env": {
100        "MCP_NAME": "web-traffic-data",
101        "PRODUCT_ID": "234",
102        "BEARER_TOKEN": "sk-12345"
103      }
104    }
105  }
106}
107```
108 
109## Development
110 
111Install dependencies:
112```bash
113npm install
114```
115 
116Build the server:
117```bash
118npm run build
119```
120 
121For development with auto-rebuild:
122```bash
123npm run watch
124```
125 
126### Debugging
127 
128Since MCP servers communicate over stdio, debugging can be challenging. We recommend using the [MCP Inspector](https://github.com/modelcontextprotocol/inspector), which is available as a package script:
129 
130```bash
131npm run inspector
132```
133 
134The Inspector will provide a URL to access debugging tools in your browser.
135 
136### Acknowledgements
137 
138- Obviously the modelcontextprotocol and Anthropic teams for the MCP Specification and integration into Claude Desktop. [https://modelcontextprotocol.io/introduction](https://modelcontextprotocol.io/introduction)
139