chore: update readme

README.md

@@ -16,24 +16,26 @@
 
 ## 项目简介
 
-追创作是一款面向内容创作者的人工智能助手，把原本分散在素材下载、字幕处理、资料检索、桌面操作与移动沟通中的工作，收束到同一个入口。它不只是工具的组合，更是一位能够理解上下文、持续推进任务的创作助手：从素材采集、字幕处理、转码导出，到资料研究、灵感整理与跨设备跟进，创作者日常反复切换的一整套流程，如今可以在同一套系统里连续完成。
+追创作是一款面向内容创作者的人工智能助手，把素材下载、字幕处理、资料检索、桌面执行与移动跟进收进同一套工作流。它不是零散工具的堆叠，而是一位能够理解上下文、持续推进任务的创作助手。
 
 ## 覆盖每一个创作现场
 
-- 🎬 创作准备更集中：素材、字幕、资料与灵感被收束到同一工作台，创作前期最分散的准备工作，可以更快沉淀为可复用的资源与更清晰的方向。
-- 🖥️ 桌面执行更完整：在用户授权下，助手不止能够理解需求，还能访问网络与本地电脑，调用工具、技能与记忆系统，把对话、操作与执行合并成同一条工作流。
-- 📱 离开桌面也不中断：强大的多渠道接入能力，让同一位助手不被桌面限制。当前支持 Telegram 等渠道接入（飞书 / 微信开发中），离开电脑后，任务依然可以继续推进。
+从创作准备、桌面处理到离开工位后的持续跟进，工作不必在不同工具和设备之间反复断开。
+
+- 🎬 创作准备：素材、字幕、资料与灵感线索可以先沉淀到同一处，再进入后续处理。
+- 🖥️ 桌面执行：在授权范围内，助手可以访问网页、调用工具、操作电脑，把对话直接推进为动作。
+- 📱 移动跟进：同一位助手可延伸到 Telegram 等渠道，离开桌面后任务也能继续推进。
 
 ## 核心能力
 
-- 📥 视频下载：千余个视频网站素材一键下载，支持身份认证，无论是 YouTube 的 BGM 还是 Bilibili 的 4K 视频，都可以沉淀为你的资源库。
-- 📝 字幕校对与翻译：已有字幕不必推倒重来，校对、翻译、复核一套衔接完成，让跨语言发布更稳，也让内容复用更高效。
-- 🎞️ 视频转码与字幕烧录：下载、翻译、转码、字幕内嵌可串成一条流水线，原本分散的发布前处理，现在可以一键交付。
-- 💡 对话式资料检索与灵感激发：资料查找、信息整理、观点延展都能在对话里完成，从灵感萌发到结构成形不必跳出当前工作流。
-- 🤖 可执行的人工智能助手：不止于回答问题，在授权范围内还可访问网页、操作电脑、调用工具，把想法直接推进为结果。
-- 🧩 多助手切换机制：不同场景可定义不同助手，分别承载各自的角色、记忆与能力边界，让研究、处理与发布互不干扰。
-- ⚙️ 人工智能原生自动化：重复性的下载、整理、处理与定时任务可以持续运行，把有限精力留给判断、审美与创作本身。
-- 📲 移动端渠道接入：同一套能力不被桌面绑定，当前支持 Telegram 等渠道接入（飞书 / 微信开发中），手机上也能持续跟进任务。
+- 📥 视频下载：从公开视频素材到需要身份认证的来源，都可以沉淀到同一资源库里。
+- 📝 字幕校对与翻译：已有字幕不必推倒重来，校对、翻译、复核可以接成一条链路。
+- 🎞️ 视频转码与字幕烧录：下载、翻译、转码、字幕内嵌可以连续完成。
+- 💡 对话式资料检索与灵感激发：资料查找、信息整理、观点延展都可以留在同一条对话里。
+- 🤖 可执行的人工智能助手：不止回答问题，在授权范围内还能访问网页、调用工具、操作电脑。
+- 🧩 多助手切换机制：不同创作场景可以用不同助手承载各自的角色、记忆与能力边界。
+- ⚙️ 人工智能原生自动化：重复性的下载、整理、处理与定时任务可以长期运行。
+- 📲 移动端渠道接入：同一套能力不被桌面绑定，手机上也能继续跟进任务。
 
 ## 产品界面
 

README_en.md

@@ -16,24 +16,26 @@
 
 ## Overview
 
-Dream Creator is an AI assistant for content creators that brings together work usually scattered across media downloads, subtitle workflows, research, desktop operations, and mobile channels. It is not just a collection of tools, but a creative assistant that understands context and keeps work moving: from sourcing media, refining subtitles, and preparing exports to researching references, shaping ideas, and following through across devices, the workflows creators switch between every day can now continue inside one system.
+Dream Creator is an AI assistant for content creators that brings media download, subtitle work, research, desktop execution, and mobile follow-up into one workflow. It is not a loose bundle of tools, but a context-aware creative assistant that keeps work moving from one step to the next.
 
 ## Built for Every Creative Context
 
-- 🎬 Creative preparation in one place: source material, subtitles, research, and idea development come together in one workspace, so the most fragmented part of creation turns into reusable assets and clearer direction faster.
-- 🖥️ Complete desktop execution: with user authorization, the assistant goes beyond understanding requests to accessing the web and the local computer, using tools, skills, and memory to unify conversation, operation, and execution in one workflow.
-- 📱 Work continues beyond the desktop: strong multi-channel access keeps the same assistant available outside the desktop. Telegram is supported today (Feishu / WeChat in development), so work does not stop when you step away from the computer.
+From preparation and desktop execution to follow-up away from the desk, work no longer has to break across different tools and devices.
+
+- 🎬 Creative preparation: source material, subtitles, references, and idea fragments can be gathered in one place before moving into the next step.
+- 🖥️ Desktop execution: within the granted scope, the assistant can access webpages, call tools, and operate the computer so conversation can turn directly into action.
+- 📱 Mobile follow-up: the same assistant can stay available through channels such as Telegram, so tasks keep moving even away from the desktop.
 
 ## Core Capabilities
 
-- 📥 Video download: one-click downloads from more than a thousand video sites, with authenticated access supported. From a YouTube BGM reference to a 4K Bilibili clip, everything can become part of your content library.
-- 📝 Subtitle proofreading and translation: existing subtitles do not need to be rebuilt from scratch. Proofreading, translation, and review connect in one flow, making multilingual publishing more reliable and content reuse more efficient.
-- 🎞️ Video transcoding and subtitle burn-in: downloading, translation, transcoding, and subtitle embedding can be chained into one pipeline, turning fragmented release prep into a one-click delivery flow.
-- 💡 Conversational research and ideation: reference gathering, information organization, and idea expansion all happen inside the conversation, so creative momentum is not broken between inspiration and structure.
-- 🤖 Executable AI assistant: it goes beyond answering questions, with the ability to access webpages, operate the computer, and call tools within the granted scope to move ideas directly into results.
-- 🧩 Multi-assistant switching: different assistants can be defined for different contexts, each with its own role, memory, and capability boundary, so research, processing, and publishing do not interfere with each other.
-- ⚙️ AI-native automation: recurring downloading, organization, processing, and scheduled tasks can keep running on their own, leaving more attention for judgment, taste, and the creative work itself.
-- 📲 Mobile channel access: the same capability set is not confined to the desktop. Telegram is supported today, with Feishu and WeChat in development, so work can continue from your phone.
+- 📥 Video download: from public footage to authenticated sources, everything can be collected into the same library.
+- 📝 Subtitle proofreading and translation: existing subtitles do not need to be rebuilt; proofreading, translation, and review can run as one chain.
+- 🎞️ Video transcoding and subtitle burn-in: downloading, translation, transcoding, and subtitle embedding can be completed as a continuous flow.
+- 💡 Conversational research and ideation: reference lookup, information organization, and idea expansion can stay inside the same conversation.
+- 🤖 Executable AI assistant: it does more than answer questions; within the granted scope it can access webpages, call tools, and operate the computer.
+- 🧩 Multi-assistant switching: different creative contexts can use different assistants, each with its own role, memory, and capability boundary.
+- ⚙️ AI-native automation: recurring downloading, organization, processing, and scheduled tasks can keep running over time.
+- 📲 Mobile channel access: the same capability set is not tied to the desktop, so work can keep moving from your phone.
 
 ## Product Preview