Proxies the Tavily API so agents can run web searches, get direct answers and retrieve recent news articles.
接入配置
选择客户端，复制配置，再按项目文档补齐环境变量。
claude mcp add tavily-proxy --env TAVILY_API_KEY='<TAVILY_API_KEY>' -- python -m mcp_server_tavily~/.codex/config.toml
[mcp_servers.tavily-proxy]
command = "python"
args = ["-m", "mcp_server_tavily"]
env = { TAVILY_API_KEY = "<TAVILY_API_KEY>" }
# codex mcp add tavily-proxy --env TAVILY_API_KEY='<TAVILY_API_KEY>' -- python -m mcp_server_tavily~/.cursor/mcp.json
{
"mcpServers": {
"tavily-proxy": {
"command": "python",
"args": [
"-m",
"mcp_server_tavily"
],
"env": {
"TAVILY_API_KEY": "<TAVILY_API_KEY>"
}
}
}
}.vscode/mcp.json
{
"servers": {
"tavily-proxy": {
"type": "stdio",
"command": "python",
"args": [
"-m",
"mcp_server_tavily"
],
"env": {
"TAVILY_API_KEY": "<TAVILY_API_KEY>"
}
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"tavily-proxy": {
"command": "python",
"args": [
"-m",
"mcp_server_tavily"
],
"env": {
"TAVILY_API_KEY": "<TAVILY_API_KEY>"
}
}
}
}claude_desktop_config.json
{
"mcpServers": {
"tavily-proxy": {
"command": "python",
"args": [
"-m",
"mcp_server_tavily"
],
"env": {
"TAVILY_API_KEY": "<TAVILY_API_KEY>"
}
}
}
}.trae/mcp.json
{
"mcpServers": {
"tavily-proxy": {
"command": "python",
"args": [
"-m",
"mcp_server_tavily"
],
"env": {
"TAVILY_API_KEY": "<TAVILY_API_KEY>"
}
}
}
}需要的环境变量
TAVILY_API_KEY
把占位值换成你自己的密钥；不要把密钥提交到代码仓库。