{"_id":"glm-mcp","_rev":"4-1ca9cbc7003a255a3074161a463f0ea1","name":"glm-mcp","dist-tags":{"latest":"1.3.0"},"versions":{"1.1.7":{"name":"glm-mcp","version":"1.1.7","keywords":["mcp","model-context-protocol","glm","zhipu","z.ai","claude","copilot","cost-optimization"],"author":{"name":"djerok"},"license":"MIT","_id":"glm-mcp@1.1.7","maintainers":[{"name":"djerok","email":"Ericcao1010@gmail.com"}],"homepage":"https://github.com/djerok/glm-mcp#readme","bugs":{"url":"https://github.com/djerok/glm-mcp/issues"},"bin":{"glm-mcp":"src/index.js"},"dist":{"shasum":"82eb77038403350c6cac6c2707c2c12f64be1d12","tarball":"https://registry.npmjs.org/glm-mcp/-/glm-mcp-1.1.7.tgz","fileCount":10,"integrity":"sha512-s/svd4dfZPkHtVkTzOmHpmIo2NnbH6fPaLzc9Oz/Aegm1Zrllixm9E7wzUKBOjAEkXiaMn7rkfgmb2H87ARRAw==","signatures":[{"sig":"MEQCIB1x+u1dfLj14nIoYjIkqSrZEdjaAPbislDS/qj/1WmUAiBkuG7rFk9cnmHlLcK/tc0azUdsArl/NLA0lOYdyQytug==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":54205},"main":"src/index.js","type":"module","engines":{"node":">=18"},"scripts":{"smoke":"node src/smoke.js","start":"node src/index.js"},"_npmUser":{"name":"djerok","email":"Ericcao1010@gmail.com"},"repository":{"url":"git+https://github.com/djerok/glm-mcp.git","type":"git","directory":"claude/glm-mcp"},"_npmVersion":"11.9.0","description":"MCP server that delegates self-contained subtasks to the GLM (Zhipu/Z.ai) Anthropic-compatible API, so Claude Code can use GLM as a cheap, peak-aware subagent. Runs standalone via npx glm-mcp for any MCP client.","directories":{},"_nodeVersion":"25.6.1","dependencies":{"zod":"^3.23.8","@modelcontextprotocol/sdk":"^1.0.0"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/glm-mcp_1.1.7_1783310367134_0.10096085610970285","host":"s3://npm-registry-packages-npm-production"}},"1.1.8":{"name":"glm-mcp","version":"1.1.8","keywords":["mcp","model-context-protocol","glm","zhipu","z.ai","claude","copilot","cost-optimization"],"author":{"name":"djerok"},"license":"MIT","_id":"glm-mcp@1.1.8","maintainers":[{"name":"djerok","email":"Ericcao1010@gmail.com"}],"homepage":"https://github.com/djerok/glm-mcp#readme","bugs":{"url":"https://github.com/djerok/glm-mcp/issues"},"bin":{"glm-mcp":"src/index.js"},"dist":{"shasum":"d819fd36055bbc9aa1164e52e69d680bfb062cdf","tarball":"https://registry.npmjs.org/glm-mcp/-/glm-mcp-1.1.8.tgz","fileCount":10,"integrity":"sha512-tM/0v+OXBwGSmYhmqoT27i2ToybMN5rchVxUtFYxwA5SrEhsQCbTkDa7zEO0bHEDPWKWY0aKeJG5QPM3cFrbzA==","signatures":[{"sig":"MEQCIBfatRPxs4TCLwt2vfINYlUi8X0czKBx7GzFnkxubIgHAiARmQ70bHR+9KS7zBts75CZiu3NqML9ZCogpCqC/skQkg==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":54246},"main":"src/index.js","type":"module","engines":{"node":">=18"},"mcpName":"io.github.djerok/glm-mcp","scripts":{"smoke":"node src/smoke.js","start":"node src/index.js"},"_npmUser":{"name":"djerok","email":"Ericcao1010@gmail.com"},"repository":{"url":"git+https://github.com/djerok/glm-mcp.git","type":"git","directory":"claude/glm-mcp"},"_npmVersion":"11.9.0","description":"MCP server that delegates self-contained subtasks to the GLM (Zhipu/Z.ai) Anthropic-compatible API, so Claude Code can use GLM as a cheap, peak-aware subagent. Runs standalone via npx glm-mcp for any MCP client.","directories":{},"_nodeVersion":"25.6.1","dependencies":{"zod":"^3.23.8","@modelcontextprotocol/sdk":"^1.0.0"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/glm-mcp_1.1.8_1783315336499_0.19195547816502212","host":"s3://npm-registry-packages-npm-production"}},"1.2.0":{"name":"glm-mcp","version":"1.2.0","keywords":["mcp","model-context-protocol","glm","zhipu","z.ai","claude","copilot","cost-optimization"],"author":{"name":"djerok"},"license":"MIT","_id":"glm-mcp@1.2.0","maintainers":[{"name":"djerok","email":"Ericcao1010@gmail.com"}],"homepage":"https://github.com/djerok/glm-mcp#readme","bugs":{"url":"https://github.com/djerok/glm-mcp/issues"},"bin":{"glm-mcp":"src/index.js"},"dist":{"shasum":"d63149743566be0cbbc2c6aa015be95ab18236f8","tarball":"https://registry.npmjs.org/glm-mcp/-/glm-mcp-1.2.0.tgz","fileCount":10,"integrity":"sha512-QfJOcJSN2uBy1JFIrtrcy4TbG18+oEDzAJplIEFD9C1NrU0w8MHQeYrJ/5DoSnylfCmV++wFJpV36Qu6MT07hA==","signatures":[{"sig":"MEQCIDX/0mWb0lSGtsv8c3Yo6RZJdw3a4HP1NvSQQJ2KjmkaAiBFwgm//e3YTg2COGlfSCFfhQvGcD1aytFpClPuvVH2zA==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":63277},"main":"src/index.js","type":"module","engines":{"node":">=18"},"mcpName":"io.github.djerok/glm-mcp","scripts":{"smoke":"node src/smoke.js","start":"node src/index.js"},"_npmUser":{"name":"djerok","email":"Ericcao1010@gmail.com"},"repository":{"url":"git+https://github.com/djerok/glm-mcp.git","type":"git","directory":"claude/glm-mcp"},"_npmVersion":"11.9.0","description":"MCP server that delegates self-contained subtasks to the GLM (Zhipu/Z.ai) Anthropic-compatible API, so Claude Code can use GLM as a cheap, peak-aware subagent. Runs standalone via npx glm-mcp for any MCP client.","directories":{},"_nodeVersion":"25.6.1","dependencies":{"zod":"^3.23.8","@modelcontextprotocol/sdk":"^1.0.0"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/glm-mcp_1.2.0_1783415599399_0.11907745763678967","host":"s3://npm-registry-packages-npm-production"}},"1.3.0":{"name":"glm-mcp","mcpName":"io.github.djerok/glm-mcp","version":"1.3.0","type":"module","bin":{"glm-mcp":"src/index.js"},"main":"src/index.js","engines":{"node":">=18"},"scripts":{"start":"node src/index.js","smoke":"node src/smoke.js"},"dependencies":{"@modelcontextprotocol/sdk":"^1.0.0","zod":"^3.23.8"},"description":"MCP server that delegates self-contained subtasks to the GLM (Zhipu/Z.ai) Anthropic-compatible API, so Claude Code can use GLM as a cheap, peak-aware subagent. Runs standalone via npx glm-mcp for any MCP client.","repository":{"type":"git","url":"git+https://github.com/djerok/glm-mcp.git","directory":"claude/glm-mcp"},"homepage":"https://github.com/djerok/glm-mcp#readme","bugs":{"url":"https://github.com/djerok/glm-mcp/issues"},"license":"MIT","author":{"name":"djerok"},"keywords":["mcp","model-context-protocol","glm","zhipu","z.ai","claude","copilot","cost-optimization"],"_id":"glm-mcp@1.3.0","_nodeVersion":"25.6.1","_npmVersion":"11.9.0","dist":{"integrity":"sha512-UmzPhw+KCZDDydXivNETVaMhi0QQbl6NLepDq+eBPPI9qWD4JfLwD+cx1SBm7awwkIPEpVoX3FmPULWW1sTfiw==","shasum":"0ead5a0fc41b9577cb810dcca925400e5d7497a5","tarball":"https://registry.npmjs.org/glm-mcp/-/glm-mcp-1.3.0.tgz","fileCount":10,"unpackedSize":64364,"signatures":[{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEUCICf/rAFP9G9KRRenKXii2V4EVBmMi7TkPTZFluvp1Ld2AiEAkExOJm0UcjjUiiwhncHUOyiGxIn1nksmxHNGqXXu53o="}]},"_npmUser":{"name":"djerok","email":"Ericcao1010@gmail.com"},"directories":{},"maintainers":[{"name":"djerok","email":"Ericcao1010@gmail.com"}],"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/glm-mcp_1.3.0_1783921876542_0.8677966825183359"},"_hasShrinkwrap":false}},"time":{"created":"2026-07-06T03:59:26.955Z","modified":"2026-07-13T05:51:16.779Z","1.1.7":"2026-07-06T03:59:27.306Z","1.1.8":"2026-07-06T05:22:16.657Z","1.2.0":"2026-07-07T09:13:19.573Z","1.3.0":"2026-07-13T05:51:16.680Z"},"bugs":{"url":"https://github.com/djerok/glm-mcp/issues"},"author":{"name":"djerok"},"license":"MIT","homepage":"https://github.com/djerok/glm-mcp#readme","keywords":["mcp","model-context-protocol","glm","zhipu","z.ai","claude","copilot","cost-optimization"],"repository":{"type":"git","url":"git+https://github.com/djerok/glm-mcp.git","directory":"claude/glm-mcp"},"description":"MCP server that delegates self-contained subtasks to the GLM (Zhipu/Z.ai) Anthropic-compatible API, so Claude Code can use GLM as a cheap, peak-aware subagent. Runs standalone via npx glm-mcp for any MCP client.","maintainers":[{"name":"djerok","email":"Ericcao1010@gmail.com"}],"readme":"# glm-mcp\n\nGLM (Zhipu / Z.ai) as a **cheap delegate** for any AI coding agent, exposed as an\n[MCP](https://modelcontextprotocol.io) (Model Context Protocol) server over stdio.\nThis is the **standalone server** behind [`glm-mcp-claude`](https://www.npmjs.com/package/glm-mcp-claude)\nand [`glm-mcp-copilot`](https://www.npmjs.com/package/glm-mcp-copilot) — usable by any MCP client\n(Claude Desktop, Cursor, Windsurf, Glama, …).\n\n## Quickstart (any MCP client)\n\n```json\n{\n  \"mcpServers\": {\n    \"glm\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"glm-mcp\"],\n      \"env\": { \"GLM_API_KEY\": \"YOUR_ZAI_KEY\" }\n    }\n  }\n}\n```\n\n## Tools\n\n- **`glm_agent`** — run GLM as a file-accessing coding agent (read/write/edit/run) on GLM tokens.\n- **`glm_delegate`** — text-in/text-out subtask on GLM (no file access).\n- **`glm_recommend`** — free advisory: GLM vs Opus, which model, and why.\n- **`glm_status`** — free status: peak window, model picks, usage ledger, config health.\n\n## Config\n\nThe server boots and answers MCP introspection **without** an API key, but every GLM call needs\n`GLM_API_KEY` (your Z.ai / Zhipu GLM Coding Plan key — https://z.ai). See [`.env.example`](./.env.example)\nfor all tuning knobs (model picks, peak window, token cap, cost bias, …).\n\n## Full docs\n\nRepo + installers + design notes: **https://github.com/djerok/glm-mcp**\n\n## License\n\nMIT © [djerok](https://github.com/djerok)\n","readmeFilename":"README.md"}