Skip to content

AI-native knowledge base & Evernote/Obsidian alternative with instant sync & native MCP. Free on Cloudflare or self-hosted via Docker.

docs: streamline README feature sections

tianma-if466a9699127d
diff --git a/README.md b/README.md
index 8d8b64f6..dfd8c626 100644
--- a/README.md
+++ b/README.md
@@ -234,14 +234,6 @@ The Remote MCP endpoint supports the stateless `2026-07-28` protocol while retai
With MCP, EdgeEver can also connect to tools such as Notion databases and Feishu Bitable, turning scattered ideas, information, and materials from everyday notes into structured data that is easier to organize, search, and manage.
-## Bring Your Own AI Models
-
-Open **Profile → AI Integrations** to add one or more OpenAI-compatible, Anthropic Messages, or Google Gemini services with your own Base URLs and API keys. Third-party API relays are also supported. Each service can contain multiple models: discover them from the provider's model-list endpoint or enter a model ID manually. A service-level switch temporarily makes all of its models unavailable, while the workspace default selects the model used for note AI.
-
-Note AI includes six focused built-in prompts across Web, Android, and iOS: summarize, translate, polish, make concise, convert to Xiaohongshu style, and convert to X (Twitter) style. You can add custom prompts for more specialized workflows. Editors on all three platforms can also run AI directly on selected text and replace only that selection. Results stream into a reviewable draft that you can retry, refine with a follow-up instruction, append, or explicitly accept as a replacement. Translation uses a language picker whose default follows the interface language: Chinese defaults to English, while English defaults to Simplified Chinese.
-
-AI requests are sent by the EdgeEver server rather than directly by the browser or native client. Model credentials are isolated by personal workspace and encrypted before being stored. Standard deployments automatically derive an AI-specific encryption key from the existing instance authentication secret, so no additional deployment variable is required. The same AI business code runs in Cloudflare Workers and the Docker/Bun runtime.
-
## Image Compression
Image compression happens in the Web client before upload and is controlled by the **Compress note images** setting. When enabled, PNG, JPEG, WebP, and AVIF files are converted to WebP when beneficial, with the longest edge limited to `2560px`. If compression does not reduce size, the original file is kept.
@@ -250,9 +242,7 @@ EdgeEver avoids Worker-side image processing to reduce compute and image-process
## Advanced Object Storage
-The instance owner can open **Settings → Advanced → OSS object storage** to send new images and attachments to an S3-compatible service such as Alibaba Cloud OSS, Tencent COS, AWS S3, MinIO, or R2. Existing resources stay in their original store, so changing the default does not migrate or break historical attachments.
-
-Before saving third-party object-storage credentials on a Cloudflare deployment, configure the `EDGE_EVER_STORAGE_ENCRYPTION_KEY` Worker Secret with a random value of at least 32 characters. EdgeEver uses this instance-level key to encrypt the access secret stored in D1. Keep the key stable and backed up; changing it makes previously saved object-storage credentials unusable.
+The instance owner can configure S3-compatible object storage under **Settings → Advanced → OSS object storage**. Changing storage does not migrate or affect existing attachments. Cloudflare deployments must also configure an `EDGE_EVER_STORAGE_ENCRYPTION_KEY` Worker Secret of at least 32 characters.
## Migration
diff --git a/README.zh-CN.md b/README.zh-CN.md
index e60b2aa1..beb2844d 100644
--- a/README.zh-CN.md
+++ b/README.zh-CN.md
@@ -233,14 +233,6 @@ Remote MCP 端点支持无状态的 `2026-07-28` 协议,同时继续兼容现
比如让AI Agent归纳你随机记录的灵感创意、针对你的笔记做精准的人物画像、构建自己的知识图谱、自动为笔记打标签)
借助 MCP,EdgeEver 还可以与 Notion Database、飞书多维表格等工具联动,把日常笔记中零散的灵感、信息和素材沉淀到结构化数据库中,方便后续整理、检索与管理。
-## 接入自己的 AI 模型
-
-进入**个人中心 → AI 集成**,可以使用自己的 Base URL 和 API Key 添加一个或多个 OpenAI 兼容协议、Anthropic Messages 或 Google Gemini 云端服务,也支持第三方中转站。每个服务都能配置多个模型:既可以从服务的模型列表接口自动发现,也可以手动输入模型 ID。服务级开关会让该服务下的全部模型暂时不可用,工作区默认模型则决定笔记 AI 实际使用哪个模型。
-
-当前 Web、Android 与 iOS 内置 6 个常用指令:总结、翻译、润色、精炼表达、转为小红书风格与转为推特风格;更细分的场景可自行添加指令。三个平台的编辑器都可以直接处理选中内容,并只替换对应选区。模型结果会先作为可审查的流式草稿展示,用户可以重试、继续提出调整要求、追加内容,或明确接受后替换原文。翻译使用带默认值的语言下拉框:中文界面默认译为英语,英文界面默认译为简体中文。
-
-AI 请求统一由 EdgeEver 服务端发出,不会由浏览器或原生客户端直接携带模型密钥。模型凭据按个人工作区隔离并加密保存;标准部署会自动从已有的实例认证 Secret 派生 AI 专用加密密钥,不需要增加任何部署变量。Cloudflare Worker 与未来的 Docker/Bun 运行时共用同一套 AI 业务代码。
-
## 图片压缩规则
图片压缩仅在 Web 端上传前执行,由设置页的“压缩笔记内图片”开关控制。启用后,浏览器会把 PNG、JPEG、WebP、AVIF 尝试压缩为 WebP,并将最长边限制在 `2560px` 以内;如果压缩结果不比原图小,则保留原图。
@@ -249,9 +241,7 @@ Cloudflare Worker 侧执行图片处理会消耗计算/图片处理额度,因
## 高级对象存储
-实例 Owner 可以进入**设置 → 高级设置 → OSS 对象存储**,让后续上传的图片和附件写入兼容 S3 API 的阿里云 OSS、腾讯云 COS、AWS S3、MinIO 或 R2。已有资源继续保留在原存储中,因此切换默认存储不会迁移历史附件,也不会让历史附件失效。
-
-在 Cloudflare 部署中保存第三方对象存储凭据前,需要先配置一个至少 32 个字符的随机 `EDGE_EVER_STORAGE_ENCRYPTION_KEY` Worker Secret。EdgeEver 会用这个实例级密钥加密保存在 D1 中的对象存储 Secret。请保持该密钥稳定并妥善备份;更换密钥会导致之前保存的对象存储凭据无法继续使用。
+实例 Owner 可在**设置 → 高级设置 → OSS 对象存储**中配置兼容 S3 API 的对象存储。切换存储不会迁移或影响已有附件。Cloudflare 部署还需配置至少 32 个字符的 `EDGE_EVER_STORAGE_ENCRYPTION_KEY` Worker Secret。
## 导入与迁移 (Migration)