docs(app): implement AI search

docs/server/api/chat.ts

@@ -1,5 +1,6 @@
-import { streamText } from 'ai'
+import { streamText, tool } from 'ai'
 import { createWorkersAI } from 'workers-ai-provider'
+import { z } from 'zod'
 
 export default defineEventHandler(async (event) => {
   const { messages } = await readBody(event)
@@ -12,11 +13,29 @@ export default defineEventHandler(async (event) => {
       }
     : undefined
   const workersAI = createWorkersAI({ binding: hubAI(), gateway })
+  const autorag = hubAutoRAG('ui3')
 
   return streamText({
-    model: workersAI('@cf/meta/llama-3.2-3b-instruct'),
+    model: workersAI('@cf/meta/llama-3.3-70b-instruct-fp8-fast'),
     maxTokens: 10000,
-    system: 'You are a helpful assistant that can answer questions and help.',
-    messages
+    messages,
+    system: `You are a helpful assistant for Nuxt UI. Check your knowledge base before answering any questions.
+    Only respond to questions using information from tool calls.
+    if no relevant information is found in the tool calls, respond, "Sorry, I don't know."
+    Format your markdown response using the following rules:
+    - Use the vue lang for code blocks syntax highlighting.
+    - Don't use markdown headings.
+    `,
+    tools: {
+      searchDocumentation: tool({
+        description: `search the documentation for information to answer questions.`,
+        parameters: z.object({
+          question: z.string().describe('the users question')
+        }),
+        execute: async ({ question }) => {
+          return (await autorag.aiSearch({ query: question })).response
+        }
+      })
+    }
   }).toDataStreamResponse()
 })