o3-search
Perform web searches using natural language queries to find current information, troubleshoot errors, or discuss complex problems.
Instructions
An AI agent with advanced web search capabilities. Useful for finding the latest information, troubleshooting errors, and discussing ideas or design challenges. Supports natural language queries.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| input | Yes | Ask questions, search for information, or consult about complex problems in English. |
Implementation Reference
- index.ts:47-83 (handler)The handler function that executes the o3-search tool logic. It takes an input string, calls OpenAI's responses.create with web_search_preview tool and reasoning effort, then returns the output text.
async ({ input }) => { try { const response = await openai.responses.create({ model: config.model, input, tools: [ { type: "web_search_preview", search_context_size: config.searchContextSize, }, ], tool_choice: "auto", parallel_tool_calls: true, reasoning: { effort: config.reasoningEffort }, }); return { content: [ { type: "text", text: response.output_text || "No response text available.", }, ], }; } catch (error) { console.error("Error calling OpenAI API:", error); return { content: [ { type: "text", text: `Error: ${error instanceof Error ? error.message : "Unknown error occurred"}`, }, ], }; } }, ); - index.ts:40-46 (schema)Input schema for the o3-search tool using Zod: a single required string input describing the user's query.
{ input: z .string() .describe( "Ask questions, search for information, or consult about complex problems in English.", ), }, - index.ts:37-83 (registration)Registration of the 'o3-search' tool on the McpServer instance, including its name, description, input schema, and handler.
server.tool( "o3-search", `An AI agent with advanced web search capabilities. Useful for finding the latest information, troubleshooting errors, and discussing ideas or design challenges. Supports natural language queries.`, { input: z .string() .describe( "Ask questions, search for information, or consult about complex problems in English.", ), }, async ({ input }) => { try { const response = await openai.responses.create({ model: config.model, input, tools: [ { type: "web_search_preview", search_context_size: config.searchContextSize, }, ], tool_choice: "auto", parallel_tool_calls: true, reasoning: { effort: config.reasoningEffort }, }); return { content: [ { type: "text", text: response.output_text || "No response text available.", }, ], }; } catch (error) { console.error("Error calling OpenAI API:", error); return { content: [ { type: "text", text: `Error: ${error instanceof Error ? error.message : "Unknown error occurred"}`, }, ], }; } }, );