Claude Watsonx Studio快速入门
- 这是基于watsonx数据MCP检索服务器项目(https://github.com/IBM/ibm-watsonxdata-mcp-server)
- 先决条件
├── Copy .env.sample to .env and populate the config values.
├── Watsonx.data service (lite is enough, in a region of your choice)
├── A vaild data API key for a watsonx.data user.
├──Should have access to the cluster resources, catalog etc.
└── Should have access to the presto and spark engines
├── Watsonx-ai project if you want to use ibm granite models.
├── Watson-ai project key.
└── Some credit in your anthropic account if you want to use anthropic claude models.
- 编辑
.env文件并添加您的API密钥：
ANTHROPIC_API_KEY= # if you want to use claude models
# Base URL for watsonx.data service, this would be different for each region.
# needed to configure MCP tool
WATSONX_DATA_BASE_URL=https://us-south.lakehouse.cloud.ibm.com/lakehouse/api/
WATSONX_DATA_API_KEY=#This is needed to connect to watsonx.data services
WATSONX_DATA_INSTANCE_ID=
#Watsonx-ai settings need if you want to use the ibm granite models.
WATSONX_API_KEY=
WATSONX_PROJECT_ID=
#URL for watsonx ai and ml service, region specific.
WATSONX_URL=https://us-south.ml.cloud.ibm.com
#change these if you want a different login than default.
WEBUI_ADMIN_EMAIL=admin@ibm.com #default, change if you want a different user.
WEBUI_ADMIN_PASSWORD=pass123 #default, change if you want a different user.
WEBUI_ADMIN_NAME=Admin #default, change if you want a different user.- 运行：
docker组成-up-d
- 打开http://localhost:3000
- 开始新的聊天，确保根据您的配置选择您选择的型号。
- 确保选择了MCP工具，以便模型可以利用它。
- 开始和克劳德聊起你的watsonx.data吧！ 🎉