create-mf2-app/cli/templates/base-fullstack-ai/convex/tools/searchMessages.ts

The stack AI moves fast with.

import { components } from "../_generated/api";
import { createTool, fetchContextMessages } from "@convex-dev/agent";
import z from "zod";
import { embed } from "ai";
import { textEmbedding } from "../models";

export const searchMessages = createTool({
  description: "Search for messages in the thread",
  args: z.object({
    query: z.string().describe("The query to search for"),
  }),
  handler: async (ctx, { query }) => {
    return fetchContextMessages(ctx, components.agent, {
      userId: ctx.userId,
      threadId: ctx.threadId,
      messages: [{ role: "user", content: query }],
      contextOptions: {
        searchOtherThreads: !!ctx.userId,
        recentMessages: 0,
        searchOptions: {
          textSearch: true,
          vectorSearch: true,
          messageRange: { before: 0, after: 0 },
          limit: 10,
        },
      },
      getEmbedding: async (text) => {
        const e = await embed({ model: textEmbedding, value: text });
        return {
          vector: e.embedding,
          vectorModel: textEmbedding.modelId,
        };
      },
    });
  },
});