BookmarkAI

English | 中文

English

BookmarkAI is a code bookmark and project knowledge tool built for the AI-assisted coding era.

It helps developers and AI build project context together: developers can manually mark important code locations, organize groups, and write notes; AI can generate bookmark candidates through BookmarkAI Skills, pointing out code it analyzed, changed, or recommends reviewing.

AI-generated candidates are staged first and reviewed by humans before becoming official bookmarks. BookmarkAI makes key logic, change traces, and project knowledge visible, navigable, and maintainable.

Core Capabilities

• Shared project context for humans and AI: manual bookmarks and AI-generated candidates become lasting project knowledge together.
• AI bookmark candidates: AI uses BookmarkAI Skills to generate structured candidates for entry points, changed areas, risks, and review targets.
• Automatic AI batch intake: validated Skill output appears in Staged Bookmarks automatically when the IDE sees a completed import batch.
• Review before merge: AI candidates do not directly change official data. Review, edit, delete, or confirm them in Staged Bookmarks.
• Better bookmark management for complex projects: organize by feature, module, task, bug, or review topic, with sorting, location review, and direct navigation back to the corresponding code.
• Key Notes for project knowledge: use Markdown to record domain concepts, module notes, design decisions, and debugging experience.

Use Cases

Daily Coding and Source Reading

• Organize complex projects by topic While reading code or developing features, add entry functions, core logic, debugging locations, and review points as bookmarks, then organize them by feature, module, task, bug, or review topic. Later, jump directly back to the corresponding code from the BookmarkAI panel. Editor Right Click -> Add Bookmark -> Select Group

• Capture project knowledge and experience Use Key Notes to record domain terms, module explanations, design decisions, debugging experience, and review conclusions.

  The fastest workflow is to open the BookmarkAI Notes panel, turn on Follow Code Selection, and keep the note detail area in reading mode. Select a term or code fragment in the editor: if a matching note already exists, the panel opens it automatically; if not, the panel shows that selection so you can click Edit to create and fill in the note under the active note group.

  You can also create notes manually from New Note in the Notes panel, or select text in the editor and use Editor Right Click -> Add Note.

Deep Integration with Claude Code, Codex, and Other AI Coding Assistants

• Use AI to generate project-wide labels Useful when joining a new project, understanding the architecture, or rebuilding project context. Ask AI to use BookmarkAI Skill bookmarkai-full to scan the project and generate global bookmark candidates for entry points, modules, and core flows. Claude Code / Codex -> BookmarkAI Skill: bookmarkai-full -> Auto Staged Bookmarks -> Review -> Merge -> Confirm

• Use AI to generate bookmarks for a feature or workflow Useful when you want to understand a feature path, business workflow, or key call chain. Ask AI to use BookmarkAI Skill bookmarkai-workflow to focus on that flow and generate related bookmark candidates. Claude Code / Codex -> BookmarkAI Skill: bookmarkai-workflow -> Auto Staged Bookmarks -> Review -> Merge -> Confirm

• Use AI to generate review bookmarks for current code changes Useful after AI or a developer finishes a round of changes. Ask AI to use BookmarkAI Skill bookmarkai-changes to analyze current changes and mark modified areas, risk points, and recommended review locations. Claude Code / Codex -> BookmarkAI Skill: bookmarkai-changes -> Auto Staged Bookmarks -> Review -> Merge -> Confirm

BookmarkAI is not just a bookmark tool. It is a shared project context system maintained by humans and AI together.

中文

BookmarkAI 是一款面向 AI 编程时代的代码书签与项目知识管理工具。

它帮助开发者和 AI 共同建设项目上下文：开发者可以手动标记关键代码、整理分组、记录笔记；AI 可以通过 BookmarkAI Skill 生成书签候选，标出它分析过、修改过或建议 Review 的位置。

所有 AI 候选都会先进入暂存区，经过人工审查后再合并为正式书签。BookmarkAI 让项目中的关键逻辑、修改线索和知识说明变得可见、可导航、可持续维护。

核心能力

• 人和 AI 共建项目上下文：手动书签与 AI 生成候选可以一起沉淀为项目长期知识。
• AI 书签候选：AI 通过 BookmarkAI Skill 生成结构化候选，记录入口、修改点、风险点和 Review 点。
• AI Batch 自动进入暂存区：Skill 生成并验证完成的候选文件会被 IDE 自动导入到 Staged Bookmarks。
• 先审查，再合并：AI 候选不会直接进入正式数据，可在 Staged Bookmarks 中查看、修改、删除或确认。
• 更适合复杂项目的书签管理：按功能、模块、任务、Bug 或 Review 主题分组，支持排序、位置复核，并可从书签直接跳转到对应代码。
• Key Notes 项目知识库：用 Markdown 记录业务概念、模块说明、设计约定和排查经验。

使用场景

日常高效编码与源码阅读

• 按主题梳理复杂项目 阅读代码或开发功能时，把入口函数、核心逻辑、排查位置、Review 重点加入书签，并按功能、模块、任务、Bug 或 Review 主题分组管理；之后可从 BookmarkAI 面板直接跳转回对应代码。 Editor Right Click -> Add Bookmark -> Select Group

• 沉淀项目知识和经验 用 Key Notes 记录业务术语、模块说明、设计约定、排查经验和 Review 结论。

  最快捷的方式是打开 BookmarkAI 的 Notes 面板，并开启 Follow Code Selection，同时保持笔记详情区处于阅读状态。然后在代码编辑器中选中一个术语或代码片段：如果已有同名 Note，面板会自动切到对应笔记；如果还没有，面板会显示这个选词，点击 Edit 就可以在当前 Note 分组下创建并补充内容。

  需要手动创建时，也可以在 Notes 面板点击 New Note，或在编辑器中选中文本后右键选择 Add Note。

与 Claude Code / Codex 等 AI Coding 深度联动

• 使用 AI 为整个项目生成标签 适合新接手项目、梳理架构或重建项目上下文。让 AI 使用 BookmarkAI Skill bookmarkai-full 扫描项目，生成入口、模块、核心流程等全局书签候选。 Claude Code / Codex -> BookmarkAI Skill: bookmarkai-full -> Auto Staged Bookmarks -> Review -> Merge -> Confirm

• 使用 AI 为指定功能或业务流程生成书签 适合理解某个功能链路、业务流程或关键调用路径。让 AI 使用 BookmarkAI Skill bookmarkai-workflow 聚焦分析，并生成该流程相关的书签候选。 Claude Code / Codex -> BookmarkAI Skill: bookmarkai-workflow -> Auto Staged Bookmarks -> Review -> Merge -> Confirm

• 使用 AI 为本次代码改动生成 Review 书签 适合在 AI 或开发者完成一轮修改后进行复查。让 AI 使用 BookmarkAI Skill bookmarkai-changes 分析当前变更，标记修改点、风险点和建议复核的位置。 Claude Code / Codex -> BookmarkAI Skill: bookmarkai-changes -> Auto Staged Bookmarks -> Review -> Merge -> Confirm

BookmarkAI 不只是书签工具。它是人和 AI 共同维护的项目上下文系统。