Add a Custom Connector
- Open Customize or Settings → Connectors.
- Choose + → Add custom connector.
- Name it KQL Search and paste the endpoint.
- Leave OAuth Client ID and Secret empty.
- In a new chat, enable it under + → Connectors.
Search thousands of community KQL examples from Claude, Codex, GitHub Copilot, ChatGPT, or any Streamable HTTP MCP client. Your AI uses the retrieved context to create the final query.
https://www.kqlsearch.com/mcpAdd the hosted server at user scope, then confirm it appears with claude mcp list.
claude mcp add --transport http --scope user kql-search https://www.kqlsearch.com/mcpRegister the hosted server, verify it with codex mcp list, then open a new Codex session.
codex mcp add kql-search --url https://www.kqlsearch.com/mcpMCP: Add Server.kql-search.Point your client or custom harness at the endpoint. No API key or OAuth credentials are required.
{ "mcpServers": { "kql-search": { "type": "http", "url": "https://www.kqlsearch.com/mcp" } } }A successful test visibly calls the KQL Search tools and returns source URLs from the library.
search_kql_queriesFind examples using text, tables, keywords, operators, or authors.
get_kql_queryRetrieve complete or safely chunked KQL and its source metadata.
list_kql_metadataExplore available tables, keywords, operators, and authors.
KQL Search never executes KQL, accesses Microsoft tenants, or calls a server-owned AI model. Community query content is returned as untrusted reference material.