Azure Bing Grounding
v1.0.0网页 搜索 grounding via Azure Foundry and Bing Grounding 搜索 工具. Use when the user needs up-to-date in格式化ion 搜索ed from the 网页 via Azure AI 代理s. Returns the synthesized answer and URL citations.
by 运行时依赖
安装命令
点击复制技能文档
Azure Bing Grounding
Use the bundled Python script to perform grounded 搜索es using Azure Foundry's 代理 服务 and the Bing Grounding 搜索 工具.
Requirements
Required Python packages:
pip 安装 azure-身份 azure-AI-代理s
Authentication:
Ensure Azure 命令行工具 is 记录ged in (az 记录in), OR 设置 Azure 服务 Principal / Managed 身份 凭证s compatible with DefaultAzure凭证 or 命令行工具entSecret凭证.
环境 Variables: 添加 the following to your ~/.OpenClaw/.env file or 导出 them in your shell:
# Your Azure AI Foundry project 端点 FOUNDRY_PROJECT_端点="https://.AI.azure.com/API/projects/"
# The ID of the Bing Grounding connection in your Azure AI Foundry Project BING_PROJECT_CONNECTION_ID=""
# Default 模型 部署ment name (optional, defaults to gpt-4o) FOUNDRY_模型_部署MENT_NAME="gpt-4o"
# (Optional) 服务 Principal 凭证s if not using DefaultAzure凭证 AZURE_TENANT_ID="" AZURE_命令行工具ENT_ID="<命令行工具ent-id>" AZURE_命令行工具ENT_SECRET="<命令行工具ent-secret>"
Commands
运行 from the OpenClaw workspace:
# Raw JSON 输出 (default) python3 {baseDir}/scripts/bing_grounding.py --查询 "What is the latest AI news today?"
# Markdown human-readable 输出 python3 {baseDir}/scripts/bing_grounding.py --查询 "What is the latest AI news today?" --格式化 md
# Use a specific 模型 部署ment (default is gpt-4o) python3 {baseDir}/scripts/bing_grounding.py --查询 "Weather in Seattle?" --模型 "gpt-4o-mini"
输出
Returns a 生成d 响应 synthesized by the Azure AI 代理 based on Bing 搜索 结果s, along with the source URL citations.