glm-mcp — run GLM as a cheap delegate for your AI coding agent

Use the GLM model (Zhipu / Z.ai) as a ~10× cheaper delegate inside your AI coding tool. Two editions, the same GLM MCP server underneath — pick your editor:

Pick your editor

Both editions now share the same primitives — MCP tools + a GLM subagent + a PreToolUse auto-routing hook. (Copilot's hook fires on all tool calls and filters internally, since VS Code ignores hook matchers; only Claude's standalone glm-code full-GLM launcher has no Copilot analog.) Everything downstream (GLM agent loop, peak-aware routing, cost bias, token cap, usage ledger, dry-run oversight) is the same server, so behavior is identical once a tool runs.

Global vs per-project

• Claude Code installs globally by default (user-scoped MCP server + ~/.claude agent/hook) — works in every project.

• GitHub Copilot — add --global to configure every workspace at once (writes VS Code's user mcp.json + settings.json); omit it to set up just the current project:

  npx glm-mcp-copilot --global --key YOUR_ZAI_KEY

Related MCP server: Ollama MCP Server

What it does (both editions)

• Offloads well-specified coding work to GLM — glm_agent lets GLM read/write/edit/run your repo directly, on GLM tokens (~10× cheaper). Your main model orchestrates + verifies.

• Peak-aware model picks, a cost bias that keeps GLM the default, a token cap toggle, a usage ledger (proof of GLM tokens), and dry-run oversight (preview a diff before applying).

Requirements

• A Z.ai / Zhipu GLM Coding Plan API key — https://z.ai (the only paid key needed).

• Node.js ≥ 18, plus the editor (Claude Code app, or VS Code + Copilot with agent mode).

License

MIT © djerok · Canonical repo: https://github.com/djerok/glm-mcp