Configure TGlider with any MCP-compatible client.
Install Node.js 24 or newer. The simplest stdio setup can run TGlider through npx -y tglider@latest.
When possible, configure TGlider per project/workspace. This avoids starting it for repositories where JS/TS semantic tools are not useful. Most clients read a project-scoped config file: Claude Code .mcp.json (claude mcp add --scope project), Cursor .cursor/mcp.json, VS Code / GitHub Copilot .vscode/mcp.json, Gemini CLI .gemini/settings.json, OpenCode opencode.json, and Pi .mcp.json. Claude Desktop only supports one global claude_desktop_config.json with no per-project scope, so pin a workspace there by adding --workspace to its args, or use Claude Code for per-project setups.
Add this product-specific instruction to your MCP client system prompt or project instructions after connecting TGlider.
When working in TypeScript or JavaScript workspaces, prefer tglider mcp semantic tools before grep, rg, or find for code navigation and refactoring. Use TGlider for symbols, declarations, references, implementations, exports, package importers, project and dependency topology, diagnostics, callers, outgoing calls, impact analysis, and preview-first rename or refactor planning. Use shell text search only for non-code assets, files outside the loaded workspace, generated output, or after TGlider cannot load or cannot answer the question. Preload the workspace at startup with --workspace (or call load first); the load is a one-time background cost, so do not fall back to text search just because loading a large workspace seems slow.Many clients use this shape. Replace the file location with your client's recommended project or workspace config file.
{
"mcpServers": {
"tglider": {
"command": "npx",
"args": ["-y", "tglider@latest"]
}
}
}VS Code uses a different server map shape. Use this for GitHub Copilot Chat or compatible VS Code MCP flows.
{
"servers": {
"tglider": {
"type": "stdio",
"command": "npx",
"args": ["-y", "tglider@latest"]
}
}
}If your client requires HTTP, start TGlider in HTTP mode and point the client at http://localhost:5002/mcp. The /health endpoint is available on the same port.
# Start TGlider in HTTP mode
npm exec -- tglider --transport http --port 5002
# Then configure your client to connect to:
# http://localhost:5002/mcpFor clients that accept a URL in mcpServers, use this shape after starting TGlider in HTTP mode.
{
"mcpServers": {
"tglider": {
"url": "http://localhost:5002/mcp"
}
}
}For VS Code and GitHub Copilot Chat, use type: "http" and the same local endpoint.
{
"servers": {
"tglider": {
"type": "http",
"url": "http://localhost:5002/mcp"
}
}
}Add --workspace <path> to the launch arguments in the configuration above so TGlider loads your repository root (a directory, not a project file) at startup, instead of waiting for the agent to call load. With it set, server_status shows the workspace as loaded — or, on a large workspace, still loading in the background under workspaceLoading — from the first message, so the agent can wait for that load to finish instead of stopping with a "No workspace loaded." message and reaching for text search. Put the flag after -- for claude mcp add, add it to the args array in JSON and Codex TOML configs, or to the single command array for OpenCode. Prefer an absolute path — it is machine-specific, so keep it in local or user scope — or a relative path (resolved against the launch directory) when the config is shared in a project.
npx -y tglider@latest --workspace /path/to/repo
# Add --no-watch to load without starting the file watcher.