Upgrade TTS to qwen3-tts-instruct-flash with instruction control

Adds an optional instructions parameter to sendVoiceMessage so the model
can describe tone, pace and emotion in natural language. Defaults the
TTS model to qwen3-tts-instruct-flash; Chelsie voice is unchanged.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

README.md

@@ -95,8 +95,8 @@ dashScopeApiKey: ''
 imageModel: 'qwen-image-2.0'
 # 是否在生成图片右下角添加 Qwen-Image 水印
 imageWatermark: false
-# 百炼平台TTS模型
+# 百炼平台TTS模型，qwen3-tts-instruct-flash 支持 instructions 指令控制语气
-ttsModel: 'qwen-tts'
+ttsModel: 'qwen3-tts-instruct-flash'
 # Jina API Key
 jinaApiKey: ''
 # SearXNG 搜索引擎地址，如 http://127.0.0.1:8080/search 必须启用允许json格式返回

src/main/kotlin/PluginConfig.kt

@@ -56,8 +56,8 @@ object PluginConfig : AutoSavePluginConfig("Config") {
     @ValueDescription("是否在生成的图片右下角添加 Qwen-Image 水印")
     val imageWatermark: Boolean by value(false)
 
-    @ValueDescription("百炼平台TTS模型")
+    @ValueDescription("百炼平台TTS模型。qwen3-tts-instruct-flash 支持 instructions 指令控制；纯发音可用 qwen3-tts-flash 或 qwen-tts")
-    val ttsModel: String by value("qwen-tts")
+    val ttsModel: String by value("qwen3-tts-instruct-flash")
 
     @ValueDescription("Jina API Key")
     val jinaApiKey by value("")

src/main/kotlin/tools/SendVoiceMessage.kt

@@ -21,7 +21,7 @@ import kotlin.time.measureTime
 class SendVoiceMessage : BaseAgent(
     tool = Tool.function(
         name = "sendVoiceMessage",
-        description = "发送一条文本转语音消息。",
+        description = "发送一条文本转语音消息。可选传入 instructions 用自然语言指令控制语气、语速、情感等表达方式，让语音更贴合当前对话氛围。",
         parameters = Parameters.buildJsonObject {
             put("type", "object")
             putJsonObject("properties") {
@@ -29,6 +29,10 @@ class SendVoiceMessage : BaseAgent(
                     put("type", "string")
                     put("description", "语音消息文本内容")
                 }
+                putJsonObject("instructions") {
+                    put("type", "string")
+                    put("description", "可选。自然语言描述本句话的表达方式，例如\"语速较快，带有明显的上扬语调\"或\"温柔知性，语调平和\"。仅支持中英文。")
+                }
             }
             putJsonArray("required") {
                 add("content")
@@ -52,6 +56,7 @@ class SendVoiceMessage : BaseAgent(
         if (event.subject !is AudioSupported) return "当前聊天环境不支持发送语音！"
 
         val content = args.getValue("content").jsonPrimitive.content
+        val instructions = args["instructions"]?.jsonPrimitive?.content?.takeIf { it.isNotBlank() }
 
         // https://help.aliyun.com/zh/model-studio/qwen-tts
         val response = httpClient.post(API_URL) {
@@ -62,6 +67,10 @@ class SendVoiceMessage : BaseAgent(
                 putJsonObject("input") {
                     put("text", content)
                     put("voice", "Chelsie") // Chelsie（女） Cherry（女） Ethan（男） Serena（女）
+                    if (instructions != null) {
+                        put("instructions", instructions)
+                        put("optimize_instructions", true)
+                    }
                 }
             }.toString())
         }