diff --git a/enjoy/src/i18n/en.json b/enjoy/src/i18n/en.json
index 17441c2b..37d71e4c 100644
--- a/enjoy/src/i18n/en.json
+++ b/enjoy/src/i18n/en.json
@@ -21,6 +21,7 @@
       "added": "Successfully added audio",
       "removed": "Successfully removed audio",
       "notFound": "Video not found",
+      "fileTooLarge": "File too large {{ file }}",
       "failedToAdd": "Failed to add audio, {{ error }}",
       "fileNotFound": "File not found {{file}}",
       "fileNotSupported": "File not supported {{file}}",
@@ -45,6 +46,7 @@
       "added": "Successfully added video",
       "removed": "Successfully removed video",
       "notFound": "Video not found",
+      "fileTooLarge": "File too large {{ file }}",
       "failedToAdd": "Failed to add video, {{ error }}",
       "fileNotFound": "File not found {{file}}",
       "fileNotSupported": "File not supported {{file}}",
@@ -82,6 +84,7 @@
       "ttsEngine": "TTS engine",
       "ttsModel": "TTS model",
       "ttsVoice": "TTS voice",
+      "ttsBaseUrl": "TTS base URL",
       "notFound": "Conversation not found",
       "contentRequired": "Content required"
     },
diff --git a/enjoy/src/i18n/zh-CN.json b/enjoy/src/i18n/zh-CN.json
index ada25107..87480325 100644
--- a/enjoy/src/i18n/zh-CN.json
+++ b/enjoy/src/i18n/zh-CN.json
@@ -21,6 +21,7 @@
       "added": "成功添加音频",
       "removed": "成功删除音频",
       "notFound": "未找到音频",
+      "fileTooLarge": "文件太大了 {{ file }}",
       "failedToAdd": "添加音频失败, {{error}}",
       "fileNotFound": "无法访问文件 {{file}}",
       "fileNotSupported": "文件不支持 {{file}}",
@@ -45,6 +46,7 @@
       "added": "成功添加视频",
       "removed": "成功删除视频",
       "notFound": "未找到视频",
+      "fileTooLarge": "文件太大了 {{ file }}",
       "failedToAdd": "添加视频失败, {{error}}",
       "fileNotFound": "无法访问文件 {{file}}",
       "fileNotSupported": "文件不支持 {{file}}",
@@ -82,6 +84,7 @@
       "ttsEngine": "TTS 引擎",
       "ttsModel": "TTS 模型",
       "ttsVoice": "TTS 声音",
+      "ttsBaseUrl": "TTS 请求地址",
       "notFound": "未找到对话",
       "contentRequired": "对话内容不能为空"
     },
diff --git a/enjoy/src/main/db/models/audio.ts b/enjoy/src/main/db/models/audio.ts
index 95ebc4b1..07973f0d 100644
--- a/enjoy/src/main/db/models/audio.ts
+++ b/enjoy/src/main/db/models/audio.ts
@@ -29,6 +29,8 @@ import webApi from "@main/web-api";
 import { startCase } from "lodash";
 import { v5 as uuidv5 } from "uuid";
 
+const SIZE_LIMIT = 1024 * 1024 * 50; // 50MB
+
 const logger = log.scope("db/models/audio");
 @Table({
   modelName: "Audio",
@@ -240,6 +242,11 @@ export class Audio extends Model<Audio> {
       throw new Error(t("models.audio.fileNotSupported", { file: filePath }));
     }
 
+    const stats = fs.statSync(filePath);
+    if (stats.size > SIZE_LIMIT) {
+      throw new Error(t("models.audio.fileTooLarge", { file: filePath }));
+    }
+
     const md5 = await hashFile(filePath, { algo: "md5" });
 
     // Generate ID
diff --git a/enjoy/src/main/db/models/speech.ts b/enjoy/src/main/db/models/speech.ts
index 3ce0808c..30c883f3 100644
--- a/enjoy/src/main/db/models/speech.ts
+++ b/enjoy/src/main/db/models/speech.ts
@@ -161,6 +161,7 @@ export class Speech extends Model<Speech> {
       engine = "openai",
       model = "tts-1",
       voice = "alloy",
+      baseUrl,
     } = configuration || {};
 
     logger.debug("Generating speech", { engine, model, voice });
@@ -176,7 +177,9 @@ export class Speech extends Model<Speech> {
       }
       const openai = new OpenAI({
         apiKey: key,
+        baseURL: baseUrl,
       });
+      logger.debug("baseURL", openai.baseURL);
 
       const file = await openai.audio.speech.create({
         input: text,
diff --git a/enjoy/src/main/db/models/video.ts b/enjoy/src/main/db/models/video.ts
index 36645088..c2899a7a 100644
--- a/enjoy/src/main/db/models/video.ts
+++ b/enjoy/src/main/db/models/video.ts
@@ -29,6 +29,8 @@ import webApi from "@main/web-api";
 import { startCase } from "lodash";
 import { v5 as uuidv5 } from "uuid";
 
+const SIZE_LIMIT = 1024 * 1024 * 100; // 100MB
+
 const logger = log.scope("db/models/video");
 @Table({
   modelName: "Video",
@@ -262,6 +264,11 @@ export class Video extends Model<Video> {
       throw new Error(t("models.video.fileNotSupported", { file: filePath }));
     }
 
+    const stats = fs.statSync(filePath);
+    if (stats.size > SIZE_LIMIT) {
+      throw new Error(t("models.video.fileTooLarge", { file: filePath }));
+    }
+
     const md5 = await hashFile(filePath, { algo: "md5" });
 
     // Generate ID
diff --git a/enjoy/src/renderer/components/audios/audio-card.tsx b/enjoy/src/renderer/components/audios/audio-card.tsx
index 5c60afc9..b61012e6 100644
--- a/enjoy/src/renderer/components/audios/audio-card.tsx
+++ b/enjoy/src/renderer/components/audios/audio-card.tsx
@@ -2,7 +2,7 @@ import { Link } from "react-router-dom";
 import { cn } from "@renderer/lib/utils";
 
 export const AudioCard = (props: {
-  audio: ResourceType;
+  audio: AudioType;
   className?: string;
 }) => {
   const { audio, className } = props;
diff --git a/enjoy/src/renderer/components/conversations/conversation-form.tsx b/enjoy/src/renderer/components/conversations/conversation-form.tsx
index c37e761a..69db60b9 100644
--- a/enjoy/src/renderer/components/conversations/conversation-form.tsx
+++ b/enjoy/src/renderer/components/conversations/conversation-form.tsx
@@ -53,6 +53,7 @@ const conversationFormSchema = z.object({
           engine: z.enum(["openai"]).default("openai"),
           model: z.string().default("tts-1"),
           voice: z.string().optional(),
+          baseUrl: z.string().optional(),
         })
         .optional(),
     })
@@ -102,6 +103,9 @@ export const ConversationForm = (props: {
           engine: conversation.engine,
           configuration: {
             ...conversation.configuration,
+            tts: {
+              ...conversation.configuration?.tts,
+            },
           },
         }
       : {
@@ -128,6 +132,10 @@ export const ConversationForm = (props: {
       configuration.baseUrl = LLM_PROVIDERS[engine]?.baseUrl;
     }
 
+    if (!configuration.tts.baseUrl) {
+      configuration.tts.baseUrl = LLM_PROVIDERS[engine]?.baseUrl;
+    }
+
     if (conversation?.id) {
       EnjoyApp.conversations
         .update(conversation.id, {
@@ -547,6 +555,21 @@ export const ConversationForm = (props: {
                 </FormItem>
               )}
             />
+
+            <FormField
+              control={form.control}
+              name="configuration.tts.baseUrl"
+              render={({ field }) => (
+                <FormItem>
+                  <FormLabel>{t("models.conversation.ttsBaseUrl")}</FormLabel>
+                  <Input {...field} />
+                  <FormDescription>
+                    {t("models.conversation.baseUrl")}
+                  </FormDescription>
+                  <FormMessage />
+                </FormItem>
+              )}
+            />
           </div>
         </ScrollArea>
 
@@ -584,7 +607,9 @@ export const ConversationForm = (props: {
           )}
 
           <Button
-            disabled={submitting || !form.formState.isDirty}
+            disabled={
+              submitting || (conversation.id && !form.formState.isDirty)
+            }
             className="w-full h-12"
             size="lg"
             type="submit"
@@ -686,8 +711,9 @@ const conversationDefaultConfiguration = {
     maxTokens: 2048,
     presencePenalty: 0,
     frequencyPenalty: 0,
-    historyBufferSize: 10,
+    historyBufferSize: 0,
     tts: {
+      baseUrl: "",
       engine: "openai",
       model: "tts-1",
       voice: "alloy",
diff --git a/enjoy/src/renderer/components/messages/assistant-message.tsx b/enjoy/src/renderer/components/messages/assistant-message.tsx
index 0596ba5e..59cca9c6 100644
--- a/enjoy/src/renderer/components/messages/assistant-message.tsx
+++ b/enjoy/src/renderer/components/messages/assistant-message.tsx
@@ -54,6 +54,7 @@ export const AssistantMessageComponent = (props: {
         engine: configuration?.tts?.engine,
         model: configuration?.tts?.model,
         voice: configuration?.tts?.voice,
+        baseUrl: configuration?.tts?.baseUrl,
       })
       .then((speech) => {
         setSpeech(speech);
diff --git a/enjoy/src/renderer/components/whisper-model-options.tsx b/enjoy/src/renderer/components/whisper-model-options.tsx
index 6ed7189d..5ba74b99 100644
--- a/enjoy/src/renderer/components/whisper-model-options.tsx
+++ b/enjoy/src/renderer/components/whisper-model-options.tsx
@@ -104,6 +104,7 @@ export const WhisperModelOptions = () => {
       }
       if (state.state === "completed") {
         model.downloaded = true;
+        setWhisperModel(model.name);
         EnjoyApp.download.removeAllListeners();
       }
 
diff --git a/enjoy/src/renderer/pages/conversations.tsx b/enjoy/src/renderer/pages/conversations.tsx
index 5328e819..1e724e25 100644
--- a/enjoy/src/renderer/pages/conversations.tsx
+++ b/enjoy/src/renderer/pages/conversations.tsx
@@ -85,7 +85,15 @@ export default () => {
 
         {conversations.map((conversation) => (
           <Link key={conversation.id} to={`/conversations/${conversation.id}`}>
-            <div className="bg-white text-primary rounded-full w-full mb-2 p-4 hover:bg-primary hover:text-white cursor-pointer flex items-center">
+            <div
+              className="bg-white text-primary rounded-full w-full mb-2 p-4 hover:bg-primary hover:text-white cursor-pointer flex items-center"
+              style={{
+                borderLeftColor: `#${conversation.id
+                  .replaceAll("-", "")
+                  .substr(0, 6)}`,
+                borderLeftWidth: 3,
+              }}
+            >
               <div className="">
                 <MessageCircleIcon className="mr-2" />
               </div>