github NitroRCr/AIaW v1.8.10
v1.8.10 | Migrating to Nya AI

one day ago

[English | 简体中文]

Migrating to Nya AI

This project is no longer maintained. We recommend that existing users migrate to Nya AI to continue their experience.

Why the Change?

As the project evolved, AIaW's original architectural design increasingly became a bottleneck that was difficult to resolve. At the same time, we wanted to implement more advanced features, and making radical overhauls on the existing codebase was no easy task. Therefore, we chose to completely rebuild a new project using a fresh technical architecture. This allowed us to rethink many underlying implementation details, add entirely new features, and even shift the overall positioning of the project.

What's New?

If AIaW was positioned as a local-first AI chat client, Nya AI is positioned as a collaborative AI platform. While AI chat was the primary function of AIaW, in Nya AI, it is just one of many core features.

Regarding chat, Nya AI retains most of AIaW’s functionality while adding many optimizations. In other areas, Nya AI introduces several new features:

Workspaces

In AIaW, workspaces were often used to separate conversations of different topics. In Nya AI, this need is met simply by creating folders within a single workspace; you can still apply different default assistants and configurations to different directories.

Consequently, the primary role of a workspace in Nya AI is collaboration rather than categorization. Similar to collaboration platforms like Notion, you can invite your team to your workspace to browse and edit content together.

No More "Cloud Sync" Needed

AIaW was local-first, relying on Dexie Cloud to sync user data across devices. In Nya AI, data is stored with a server-centric approach, supported by the Zero sync engine to achieve real-time queries, optimistic updates, and hot data caching. This ensures consistent data across all devices natively.

To be clear, Nya AI is no longer local-first. While it achieves an interaction experience close to local-first apps thanks to the sync engine, it is essentially a server-priority architecture. Now, you must log in to use it (unauthenticated users can only browse publicly published pages), and offline capabilities are more limited—you can read recent content offline but cannot make modifications. This was a necessary trade-off: the new architecture allows for better data consistency, more complex features, and the ability to handle much larger volumes of data.

Furthermore, a highlight of the new architecture is the achievement of full data control. The Dexie Cloud extension used by AIaW is closed-source, meaning you had to use their SaaS for cloud synchronization. However, the sync engine used by Nya AI, @rocicorp/zero, is open-source. Nya AI itself is also fully open-source, and the S3 storage it relies on has many compatible open-source alternatives. This means the entire chain is open-source and self-hostable—great news for self-hosting users who want total control over their data.

Pages

This is an indispensable feature for a collaborative platform. Pages are Notion-like, WYSIWYG collaborative documents. They support:

  • Markdown Compatibility: Use Markdown syntax to type, paste Markdown, or export to Markdown.
  • Docx Support: Import from and export to .docx files.
  • Edit History: Browse and revert to previous versions at any time.
  • Integrated AI: Open an AI chat on the right side to ask questions or have the AI edit the page content.

Publishing Content

Pages, chats, and files can all be published (Right-click item in the sidebar -> Publish). Once published, they are publicly accessible (read-only) via a link. Sub-items of a published item are automatically published as well.

Search

This functions like a hybrid of traditional search engines and AI search. The left side shows search results, while the right side generates an AI-summarized answer with the ability to continue the conversation.

Channels

Similar to a group chat for all workspace members, channels can be used for communication and collaboration, or for individuals to transfer text and files across different devices.

Files

You can upload files; document-type files are automatically parsed into text for use in AI conversations. You can also reference files in pages or channels. We do not restrict the types of files you can upload—you can even use it as a cloud drive, as long as the file size stays within the limit.

Migrating Data

If you want to migrate your data from AIaW to Nya AI, please follow these steps:

  1. At the bottom of the AIaW Settings page, click User Data -> Export to export your data as a JSON file.
  2. Open and log in to Nya AI.
  3. At the bottom of the Nya AI Settings page, click Import Data and select the JSON file you just exported.
  4. Wait for the import to complete. The imported data will be saved in a folder named "AIaW Import" at the root of your current workspace.

This method works whether you are migrating from the official SaaS (aiaw.app) or a self-hosted AIaW instance to the official Nya AI (nyaai.cc) or a self-hosted Nya AI instance.

If the import fails due to network fluctuations, try the following:

  1. Refresh the page.
  2. Delete the partially imported "AIaW Import" folder.
  3. Try the import again.

What is Imported?

Conversations, messages, and assistants—along with message attachments and assistant avatars—will be imported. Workspaces will be converted into folders, maintaining their original nested structure. However, other data—including settings, service providers, and plugins—will not be imported and must be reconfigured manually.

Not Ready to Migrate?

If you are wondering if Nya AI can completely replace AIaW right now, the answer is: not quite.

On one hand, Nya AI has simplified some of AIaW’s complex but low-frequency features, such as:

  • Custom Service Providers: Many users complained that custom provider configurations in AIaW were too complex. Nya AI’s design is simpler and more intuitive but no longer supports complex routing configurations.
  • Prompt Variables: Nya AI currently does not support custom prompt variables. If you need this feature, please let us know.

On the other hand, Nya AI is currently only available as a Web version and lacks a native desktop client. This means features like cross-origin requests and local MCP are currently unavailable.

It would be easy to build a cross-platform client the same way we did for AIaW, so why haven't we?

Because I am still not satisfied with that solution (Tauri for desktop + CapacitorJS for mobile); there are still problems without perfect solutions. I will explore better cross-platform options when I have time before building a native client.

So, if you absolutely require these features, you can continue using AIaW for now. However, please note that AIaW will no longer receive functional updates or active maintenance. Nya AI is still in its early stages, and we will continue to add more features. We hope you will eventually migrate to Nya AI for a better experience!

Thank You for Your Support

Every form of support—whether it’s giving us a Star, submitting an Issue, recommending us to others, subscribing to cloud sync, or simply using AIaW—is the driving force behind this project. We sincerely thank you for your support.

The AI industry is moving incredibly fast. In the past year, we have witnessed so much: MCP, Skills, Deepseek R1, Nano banana, Claude code, OpenClaw... new concepts, protocols, and products are emerging constantly.

While keeping up with the speed of AI development isn't easy, we refuse to stand still at just being a "chatbot." This is one of the reasons we built a new project—a new starting point significantly raises the ceiling of what we can achieve. We will continue to dedicate ourselves to building even better open-source AI projects and products.


[English | 简体中文]

迁移至 Nya AI

此项目已停止维护,我们建议现有用户迁移至 Nya AI 继续使用。

为什么

随着项目的发展,AIaW 最初的架构设计愈发成为瓶颈,且难以解决。同时,我们想要更多的新功能和新特性,在现有项目基础上进行大刀阔斧的更改并不容易。因此,我们选择使用新的技术架构完全重新构建了一个新项目。这允许我们在许多底层细节实现上重新思考,添加许多全新的功能特性,甚至改变整个项目的定位。

有什么新的

如果说 AIaW 的定位是本地优先的 AI 对话客户端,那 Nya AI 的定位应该是可协作 AI 平台。AI 对话是 AIaW 的主要功能,但在 Nya AI 中,AI 对话仅仅是主要功能之一。

在对话方面,Nya AI 保留了 AIaW 的大部分功能,此外还有许多新的优化。而在其他方面,Nya AI 有许多新的功能特性:

工作区

在 AIaW 中,工作区常用来分隔不同主题的对话,但在 Nya AI 中,仅仅在单个工作区中创建不同的文件夹便可以满足这一需求,你同样可以为不同的目录应用不同的默认助手等配置。

因此在 Nya AI 中,工作区的主要作用是协作而不是划分。是的,与 Notion 等协作平台类似,你可以邀请你的团队加入你的工作区,一同浏览、编辑工作区的内容。

不再需要云同步

AIaW 是本地优先的,依靠 Dexie Cloud 在不同设备之间同步用户数据。但在 Nya AI,数据是以服务端为中心储存的,辅以同步引擎 Zero 实现实时查询、乐观更新和热数据缓存,本身就能够在不同设备展示一致的数据。是的,Nya AI 不再是本地优先的了,虽然依靠同步引擎达到了接近本地优先应用的交互体验,但这本质上仍然是服务端优先的架构。现在,登录后才能使用,不登录只能浏览他人公开发布的页面;离线能力也更有限,只能离线读取最近的部分内容,无法离线修改。但这是必要的取舍,新架构能够让我们获得更好的数据一致性,实现更复杂的功能,以及承载更大的数据量。

此外,新架构的一大亮点是,实现了对数据的完全控制。AIaW 依赖的 Dexie Cloud 拓展是闭源的,意味着你必须使用他的 SaaS 才能实现云同步。但 Nya AI 使用的同步引擎 @rocicorp/zero 是开源的,同时 Nya AI 本身也是完全开源的,依赖的 S3 储存也有很多兼容的开源替代。这意味着整个链条都是开源可自部署的。这对自部署用户来说是一个好消息,这允许你完全控制自部署站点的所有数据。

页面

这应该是协作平台不可缺少的功能。页面是类 Notion 的所见即所得、可协作的文档。支持:

  • 兼容 Markdown:使用 Markdown 语法输入、粘贴 Markdown、导出为 Markdown
  • 从 Docx 文件导入 / 导出为 Docx
  • 编辑记录:可随时浏览、退回到历史版本
  • 可在右侧打开 AI 对话,向 AI 提问或是让 AI 编辑页面

发布内容

页面、对话、文件等都可以发布(在右侧边栏右键项目->发布),发布后可通过该链接公开访问(只读),发布的项目的子项目也会随之发布。

搜索

类似于传统搜索引擎和 AI 搜索的结合,左侧是搜索结果,右侧会生成 AI 总结回答,也可进行后续对话。

频道

类似于工作区所有成员的群聊,可用于交流、协作,也可用于个人在不同设备间传文本、文件等。

文件

你可以上传文件,文档类型的文件会自动解析为文本,以便在 AI 对话中使用,你也可以在页面/频道中引用文件。我们并没有限制可上传的文件类型,你甚至可以当作云盘使用,只要文件大小不超过上限即可。

迁移数据

如果你想将 AIaW 中的数据迁移至 Nya AI,请进行如下操作:

  1. 在 AIaW 设置页面的底部,点击 用户数据 -> 导出 来导出数据,导出的数据为一个 JSON 文件
  2. 打开并登录 Nya AI
  3. 在 Nya AI 设置页面的底部,点击 导入数据,选择刚才导出的 JSON 文件
  4. 等待导入完成。导入的数据会保存在当前工作区根目录的 “AIaW 导入” 文件夹中

无论是从官方 SaaS (aiaw.app) 还是自部署的 AIaW 迁移至 官方 SaaS (nyaai.cc) 或自部署的 Nya AI,都可以按上述方法迁移数据。

如果因为网络波动等原因在导入过程中失败,可按以下步骤重试:

  1. 刷新页面
  2. 删除导入到一半的 “AIaW 导入” 文件夹
  3. 重新导入

导入的数据

对话、消息、助手,以及消息中的附件、助手的头像等数据会被导入。工作区会被导入为文件夹,并保留其原有的嵌套结构。但其他数据,包括设置、服务商、插件等,不会导入,需要手动迁移。

不迁移?

如果你问 Nya AI 能否完全替代 AIaW,那么答案是,不能。

一方面,Nya AI 简化了一些 AIaW 中复杂但使用频率较低的功能,比如:

  • 自定义服务商:有不少用户抱怨 AIaW 中的自定义服务商配置过于复杂难懂,因此 Nya AI 中的自定义服务商的设计更加简单易懂,但不再支持复杂的路由配置。
  • 提示词变量:目前 Nya AI 暂不支持自定义提示词变量,如果你需要这个功能,请让我们知道

另一方面,Nya AI 目前只有 Web 版,没有本地客户端,这意味跨域、本地 MCP 等功能目前无法实现。

以 AIaW 相同的方式构建跨平台客户端应当很容易,但我们并没有这么做,为什么?

因为我对这套方案(桌面端 tauri + 移动端 capacitorjs)仍然不够满意,仍有一些问题没有完美的解决方案。我会在有空的时候探索更好的跨平台方案,之后再构建本地客户端。

所以,如果你确实需要这些功能,那么可以暂时继续使用 AIaW,不过 AIaW 将不再得到功能更新和积极维护。此外,Nya AI 仍处于早期,我们会不断地添加更多功能。因此,我们还是希望您最终能够迁移至 Nya AI 以获得更好的使用体验!

感谢您的支持

您以何种形式对我们的支持,无论是给我们 star、提出 issue、向他人推荐、为我们付费、或仅仅是使用 AIaW,都是项目发展的动力。我们由衷地感谢您的支持。

AI 行业的发展十分迅速,这一年有余,我们也见证了许多。MCP、Skills、Deepseek R1、Nano banana、Claude code、OpenClaw...各种新的概念、协议、产品层出不穷。

尽管要跟上 AI 的发展速度并不容易,但我们绝不能在 chatbot 上固步自封。这也是我们构建新项目的原因之一,新的起点也显著提高了我们能够到达的上限。我们会继续致力于构建更优秀的开源 AI 项目/产品。

Full Changelog: v1.8.9...v1.8.10

Don't miss a new AIaW release

NewReleases is sending notifications on new releases.