feat(agent): add voice input and photo attachments to assistant

Wire mic through Whisper-compatible transcriptions on LLM_BASE_URL.

Photos upload to MinIO and reach the vision model as base64 image_url parts.

src/lib/llm/transcribe.ts

@@ -0,0 +1,47 @@
+import { getLlmConfig } from "./config";
+
+const MAX_AUDIO_BYTES = 25 * 1024 * 1024;
+
+export async function transcribeAudioFile(file: File | Blob, filename: string): Promise<string> {
+  if (file.size > MAX_AUDIO_BYTES) {
+    throw new Error("Recording exceeds 25 MB limit");
+  }
+
+  const config = getLlmConfig();
+  if (config.provider === "mock" || !config.baseUrl) {
+    return "add milk to the shopping list";
+  }
+
+  const formData = new FormData();
+  formData.append("file", file, filename);
+  formData.append("model", "whisper-1");
+
+  const url = `${config.baseUrl.replace(/\/$/, "")}/audio/transcriptions`;
+  const headers: Record<string, string> = {};
+  if (config.apiKey) {
+    headers.Authorization = `Bearer ${config.apiKey}`;
+  }
+
+  const response = await fetch(url, {
+    method: "POST",
+    headers,
+    body: formData,
+  });
+
+  if (!response.ok) {
+    const detail = await response.text();
+    throw new Error(`Transcription failed (${response.status}): ${detail.slice(0, 400)}`);
+  }
+
+  const contentType = response.headers.get("content-type") ?? "";
+  if (contentType.includes("application/json")) {
+    const payload = (await response.json()) as { text?: string };
+    const text = payload.text?.trim();
+    if (!text) throw new Error("Transcription returned empty text");
+    return text;
+  }
+
+  const text = (await response.text()).trim();
+  if (!text) throw new Error("Transcription returned empty text");
+  return text;
+}