Skip to main content
Glama
evals.ts1.66 kB
//evals.ts import { EvalConfig } from 'mcp-evals'; import { openai } from "@ai-sdk/openai"; import { grade, EvalFunction } from "mcp-evals"; const search_googleEval: EvalFunction = { name: 'search_google Tool Evaluation', description: 'Evaluates the functionality of the search_google tool', run: async () => { const result = await grade(openai("gpt-4"), "Perform a Google search for 'SearchAPI.site' using the 'search_google' tool and provide the first few search results with their titles, snippets, and links."); return JSON.parse(result); } }; const search_google_imagesEval: EvalFunction = { name: 'search_google_images Evaluation', description: 'Evaluates the google images search functionality', run: async () => { const result = await grade(openai("gpt-4"), "Please use search_google_images to find images of corgis wearing hats. Include titles, thumbnails, and source links in the response."); return JSON.parse(result); } }; const search_youtubeEval: EvalFunction = { name: 'search_youtube Tool Evaluation', description: 'Evaluates the YouTube search tool functionality', run: async () => { const result = await grade(openai("gpt-4"), "Please search YouTube for the latest AI advancements and provide the video titles, thumbnails, descriptions, and links."); return JSON.parse(result); } }; const config: EvalConfig = { model: openai("gpt-4"), evals: [search_googleEval, search_google_imagesEval, search_youtubeEval] }; export default config; export const evals = [search_googleEval, search_google_imagesEval, search_youtubeEval];

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/mrgoonie/searchapi-mcp-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server