{"_id":"@edjl/llm-mcp","_rev":"2-1dae04a8e6e3c7832d9657e07638a128","name":"@edjl/llm-mcp","dist-tags":{"latest":"1.0.8"},"versions":{"1.0.7":{"name":"@edjl/llm-mcp","version":"1.0.7","keywords":["mcp","llm","openai","gemini","ai","querier"],"author":{"name":"Eduardo Lorenzo","email":"contact@edjl.dev"},"license":"MIT","_id":"@edjl/llm-mcp@1.0.7","maintainers":[{"name":"portujua","email":"ejlorenzo19@gmail.com"}],"bin":{"llm-mcp":"build/index.js"},"dist":{"shasum":"2cae122dbabc8f461fc6d3444a3bdf274524ec74","tarball":"https://registry.npmjs.org/@edjl/llm-mcp/-/llm-mcp-1.0.7.tgz","fileCount":22,"integrity":"sha512-vrXpxMDjvC/JPPgy2bAzysXYGTR4VODfSJADs6RaPV+7fHY4SikLXPsRCxfEMjvO81XvxVf/VrPgGMXFrp8euQ==","signatures":[{"sig":"MEYCIQD+b03smSoE3Zug00zPhbYfgRREXEykHNUvbDOocs/brAIhAOpbX9FF6ar7izL9hmesFzi/iEnXeoddCM089wp+XARw","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":29350},"main":"build/index.js","type":"module","types":"./build/index.d.ts","gitHead":"4432f791bfada1420f99d9e25f3120916dad588d","scripts":{"dev":"tsc --watch","build":"tsc","clean":"rm -rf build","start":"node build/index.js","prebuild":"npm run clean","postbuild":"chmod +x build/index.js"},"_npmUser":{"name":"portujua","email":"ejlorenzo19@gmail.com"},"_npmVersion":"10.9.2","description":"MCP server for querying LLMs (OpenAI and Google Gemini) using llm-querier","directories":{},"_nodeVersion":"22.17.1","dependencies":{"zod":"^3.24.1","llm-querier":"^1.0.2","@modelcontextprotocol/sdk":"^1.13.2"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"devDependencies":{"typescript":"^5.7.2","@types/node":"^22.10.2"},"_npmOperationalInternal":{"tmp":"tmp/llm-mcp_1.0.7_1753463976906_0.4725044314007689","host":"s3://npm-registry-packages-npm-production"}},"1.0.8":{"name":"@edjl/llm-mcp","version":"1.0.8","description":"MCP server for querying LLMs (OpenAI and Google Gemini) using llm-querier","main":"build/index.js","type":"module","bin":{"llm-mcp":"build/index.js"},"scripts":{"build":"tsc","clean":"rm -rf build","prebuild":"npm run clean","postbuild":"chmod +x build/index.js","start":"node build/index.js","dev":"tsc --watch"},"keywords":["mcp","llm","openai","gemini","ai","querier"],"author":{"name":"Eduardo Lorenzo","email":"contact@edjl.dev"},"license":"MIT","dependencies":{"@modelcontextprotocol/sdk":"^1.13.2","llm-querier":"^1.0.2","zod":"^3.24.1"},"devDependencies":{"@types/node":"^22.10.2","typescript":"^5.7.2"},"publishConfig":{"access":"public"},"_id":"@edjl/llm-mcp@1.0.8","gitHead":"f406322d878ed03679d7958181fa3e7f7ccf845a","types":"./build/index.d.ts","_nodeVersion":"22.21.1","_npmVersion":"10.9.4","dist":{"integrity":"sha512-UAd8MM9Nsf76b4Syw+lExQfCaPPG2sfd15ev9Z3rGHU8OSOuAlSc2aZ9sEmVzq/USjFFI7UC5afIZIe7EFrDJA==","shasum":"c85ecccda4c7e2f0e3d4c3bfbdbe7dbf502d676c","tarball":"https://registry.npmjs.org/@edjl/llm-mcp/-/llm-mcp-1.0.8.tgz","fileCount":22,"unpackedSize":29350,"signatures":[{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEYCIQDkA/ObuhdLQFIP7WJEHgCiMv3Kjzv1vcNLfYolh/7vLQIhALtiN84M+Mq4uElS2MGFVlcsjFE9m+Pf4UbY/tr4CYUz"}]},"_npmUser":{"name":"portujua","email":"ejlorenzo19@gmail.com"},"directories":{},"maintainers":[{"name":"portujua","email":"ejlorenzo19@gmail.com"}],"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/llm-mcp_1.0.8_1764704239698_0.05844822053014331"},"_hasShrinkwrap":false}},"time":{"created":"2025-07-25T17:19:36.802Z","modified":"2025-12-02T19:37:20.110Z","1.0.7":"2025-07-25T17:19:37.092Z","1.0.8":"2025-12-02T19:37:19.901Z"},"author":{"name":"Eduardo Lorenzo","email":"contact@edjl.dev"},"license":"MIT","keywords":["mcp","llm","openai","gemini","ai","querier"],"description":"MCP server for querying LLMs (OpenAI and Google Gemini) using llm-querier","maintainers":[{"name":"portujua","email":"ejlorenzo19@gmail.com"}],"readme":"# LLM MCP Server\n\nA Model Context Protocol (MCP) server that provides tools to query Large Language Models (LLMs) using the llm-querier library. Currently supports OpenAI and Google Gemini.\n\n## Installation\n\n### Option A: Global Installation\n```bash\nnpm install -g @edjl/llm-mcp\n```\n\n### Option B: Use with npx (no installation required)\n```bash\nnpx -y @edjl/llm-mcp\n```\n\n## Configuration\n\nSet the following environment variables:\n\n### OpenAI Configuration\n- `OPENAI_API_KEY`: Your OpenAI API key (required for OpenAI tool)\n- `OPENAI_MODEL`: OpenAI model to use (default: `o3`)\n\n### Google Gemini Configuration\n- `GEMINI_API_KEY`: Your Google Gemini API key (required for Gemini tool)\n- `GEMINI_MODEL`: Gemini model to use (default: `gemini-2.5-pro`)\n\nNote: You can configure just one provider or both. The server will only enable tools for configured providers.\n\n## Available Tools\n\n### `llm_ask_openai`\nAsk a single query prompt to OpenAI. Provide as much context as possible. This is a single call - no conversation state is maintained.\n\nParameters:\n- `prompt` (required): The query prompt to send to OpenAI\n- `context` (optional): Array of additional context strings to enhance the prompt\n- `examples` (optional): Array of examples to guide the response\n- `images` (optional): Array of image URLs or base64 encoded images\n- `scrapeUrls` (optional): Array of URLs to scrape and include as context\n- `fileUrls` (optional): Array of file URLs to download and include as context\n\n### `llm_ask_gemini`\nAsk a single query prompt to Google Gemini. Provide as much context as possible. This is a single call - no conversation state is maintained.\n\nParameters:\n- `prompt` (required): The query prompt to send to Google Gemini\n- `context` (optional): Array of additional context strings to enhance the prompt\n- `examples` (optional): Array of examples to guide the response\n- `images` (optional): Array of image URLs or base64 encoded images\n- `videos` (optional): Array of video URLs (Google AI supports video input)\n- `scrapeUrls` (optional): Array of URLs to scrape and include as context\n- `fileUrls` (optional): Array of file URLs to download and include as context\n\n## Usage with Cursor\n\nAdd to your Cursor settings:\n\n### Option A: Global Installation\n```json\n{\n  \"mcpServers\": {\n    \"llm-mcp\": {\n      \"command\": \"llm-mcp\",\n      \"env\": {\n        \"OPENAI_API_KEY\": \"your-openai-api-key\",\n        \"OPENAI_MODEL\": \"o3\",\n        \"GEMINI_API_KEY\": \"your-gemini-api-key\",\n        \"GEMINI_MODEL\": \"gemini-2.5-pro\"\n      }\n    }\n  }\n}\n```\n\n### Option B: Using npx\n```json\n{\n  \"mcpServers\": {\n    \"llm-mcp\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@edjl/llm-mcp\"],\n      \"env\": {\n        \"OPENAI_API_KEY\": \"your-openai-api-key\",\n        \"OPENAI_MODEL\": \"o3\",\n        \"GEMINI_API_KEY\": \"your-gemini-api-key\",\n        \"GEMINI_MODEL\": \"gemini-2.5-pro\"\n      }\n    }\n  }\n}\n```\n\n## Examples\n\n### Basic Query to OpenAI\n```javascript\nconst result = await use_mcp_tool({\n  server_name: \"llm-mcp\",\n  tool_name: \"llm_ask_openai\",\n  arguments: {\n    prompt: \"Explain the concept of quantum computing in simple terms\"\n  }\n});\n```\n\n### Query with Context and Examples\n```javascript\nconst result = await use_mcp_tool({\n  server_name: \"llm-mcp\",\n  tool_name: \"llm_ask_gemini\",\n  arguments: {\n    prompt: \"Write a haiku about programming\",\n    context: [\"Focus on the debugging process\", \"Make it humorous\"],\n    examples: [\"Bugs hide in the code / Like ninjas in the shadows / Coffee is my sword\"]\n  }\n});\n```\n\n### Query with Web Scraping\n```javascript\nconst result = await use_mcp_tool({\n  server_name: \"llm-mcp\",\n  tool_name: \"llm_ask_openai\",\n  arguments: {\n    prompt: \"Summarize the main points from this article\",\n    scrapeUrls: [\"https://example.com/article\"]\n  }\n});\n```\n\n### Query with Images (Vision Models)\n```javascript\nconst result = await use_mcp_tool({\n  server_name: \"llm-mcp\",\n  tool_name: \"llm_ask_gemini\",\n  arguments: {\n    prompt: \"What's in this image?\",\n    images: [\"https://example.com/image.jpg\"]\n  }\n});\n```\n\n## Notes\n\n- This MCP server uses the `llm-querier` library under the hood\n- Each query is independent - no conversation history is maintained\n- The server only loads tools for providers that have API keys configured\n- For more advanced usage, refer to the llm-querier documentation\n\n## License\n\nMIT","readmeFilename":"README.md"}