diff --git a/electron-builder.yml b/electron-builder.yml
index 87128027..43ee4afb 100644
--- a/electron-builder.yml
+++ b/electron-builder.yml
@@ -56,5 +56,4 @@ electronDownload:
 afterSign: scripts/notarize.js
 releaseInfo:
   releaseNotes: |
-    支持保存自定义智能体
-    修复话题重命名的问题
+    错误修复，优化体验
diff --git a/src/renderer/src/i18n/index.ts b/src/renderer/src/i18n/index.ts
index 4f3496d2..95cf273c 100644
--- a/src/renderer/src/i18n/index.ts
+++ b/src/renderer/src/i18n/index.ts
@@ -212,8 +212,7 @@ const resources = {
         'keep_alive_time.description': 'The time in minutes to keep the connection alive, default is 5 minutes.'
       },
       error: {
-        'chat.response':
-          'Something went wrong. Please check your API key in settings > providers, or check your proxy in Settings > General > Proxy'
+        'chat.response': 'Something went wrong. Please check if you have set your API key in the Settings > Providers'
       }
     }
   },
@@ -427,7 +426,7 @@ const resources = {
         'keep_alive_time.description': '对话后模型在内存中保持的时间（默认：5分钟）'
       },
       error: {
-        'chat.response': '出错了，这通常是由于没有正确配置 API 密钥或代理导致的, 请前往设置 > 模型提供商中配置 API 密钥'
+        'chat.response': '出错了，如果没有配置 API 密钥，请前往设置 > 模型提供商中配置密钥'
       }
     }
   }
diff --git a/src/renderer/src/pages/home/components/Messages.tsx b/src/renderer/src/pages/home/components/Messages.tsx
index 4cbbf377..c744140a 100644
--- a/src/renderer/src/pages/home/components/Messages.tsx
+++ b/src/renderer/src/pages/home/components/Messages.tsx
@@ -2,7 +2,7 @@ import { useAssistant } from '@renderer/hooks/useAssistant'
 import { useProviderByAssistant } from '@renderer/hooks/useProvider'
 import { fetchChatCompletion, fetchMessagesSummary } from '@renderer/services/api'
 import { EVENT_NAMES, EventEmitter } from '@renderer/services/event'
-import { estimateHistoryTokenCount, filterAtMessages } from '@renderer/services/messages'
+import { estimateHistoryTokenCount, filterMessages } from '@renderer/services/messages'
 import LocalStorage from '@renderer/services/storage'
 import { Assistant, Message, Model, Topic } from '@renderer/types'
 import { getBriefInfo, runAsyncFunction, uuid } from '@renderer/utils'
@@ -77,7 +77,7 @@ const Messages: FC<Props> = ({ assistant, topic }) => {
         setTimeout(() => EventEmitter.emit(EVENT_NAMES.AI_AUTO_RENAME), 100)
       }),
       EventEmitter.on(EVENT_NAMES.REGENERATE_MESSAGE, async (model: Model) => {
-        const lastUserMessage = last(filterAtMessages(messages).filter((m) => m.role === 'user'))
+        const lastUserMessage = last(filterMessages(messages).filter((m) => m.role === 'user'))
         if (lastUserMessage) {
           const content = `[@${model.name}](#)  ${getBriefInfo(lastUserMessage.content)}`
           onSendMessage({ ...lastUserMessage, id: uuid(), type: '@', content })
diff --git a/src/renderer/src/services/api.ts b/src/renderer/src/services/api.ts
index 3c198a85..9a61753e 100644
--- a/src/renderer/src/services/api.ts
+++ b/src/renderer/src/services/api.ts
@@ -14,7 +14,7 @@ import {
   getTranslateModel
 } from './assistant'
 import { EVENT_NAMES, EventEmitter } from './event'
-import { filterAtMessages } from './messages'
+import { filterMessages } from './messages'
 import ProviderSDK from './ProviderSDK'
 
 export async function fetchChatCompletion({
@@ -61,11 +61,12 @@ export async function fetchChatCompletion({
   }, 1000)
 
   try {
-    await providerSdk.completions(filterAtMessages(messages), assistant, ({ text, usage }) => {
+    await providerSdk.completions(filterMessages(messages), assistant, ({ text, usage }) => {
       message.content = message.content + text || ''
       message.usage = usage
       onResponse({ ...message, status: 'pending' })
     })
+    message.status = 'success'
   } catch (error: any) {
     message.content = `Error: ${error.message}`
     message.status = 'error'
diff --git a/src/renderer/src/services/messages.ts b/src/renderer/src/services/messages.ts
index 50497872..6c662045 100644
--- a/src/renderer/src/services/messages.ts
+++ b/src/renderer/src/services/messages.ts
@@ -1,11 +1,11 @@
 import { Assistant, Message } from '@renderer/types'
 import { GPTTokens } from 'gpt-tokens'
-import { takeRight } from 'lodash'
+import { isEmpty, takeRight } from 'lodash'
 
 import { getAssistantSettings } from './assistant'
 
-export const filterAtMessages = (messages: Message[]) => {
-  return messages.filter((message) => message.type !== '@')
+export const filterMessages = (messages: Message[]) => {
+  return messages.filter((message) => message.type !== '@').filter((message) => !isEmpty(message.content.trim()))
 }
 
 export function estimateInputTokenCount(text: string) {
@@ -24,7 +24,7 @@ export function estimateHistoryTokenCount(assistant: Assistant, msgs: Message[])
     model: 'gpt-4o',
     messages: [
       { role: 'system', content: assistant.prompt },
-      ...filterAtMessages(takeRight(msgs, contextCount)).map((message) => ({
+      ...filterMessages(takeRight(msgs, contextCount)).map((message) => ({
         role: message.role,
         content: message.content
       }))