🧠 AI-Powered Personal Knowledge Brain 🧠 AI 驱动的个人知识大脑

Every File You Save.
Automatically Understood.
你存下的每一份文件,
都被 AI 自动理解。

AiSyncMe ingests your PDFs, docs, voice memos, images, and web pages — then uses AI to summarize, tag, and connect them into a living knowledge graph you can actually talk to. AiSyncMe 摄入你的 PDF、文档、语音备忘录、图片和网页,用 AI 自动生成摘要、打标签、建立关联,构成一个你可以直接"对话"的活知识图谱。

AiSyncMe — Knowledge Graph AiSyncMe — 知识图谱
RECENT最近入库
📄 Q2 Strategy.pdf
🎙️ Team Standup.m4a
🌐 Product Research产品调研
Strategy
📄
Q2 Report
🎙️
Standup
🌐
Research
🖼️
Mockup
Notes
Plan
KPI
30+
Supported File Formats
支持文件格式
0
Manual Tags Required
需要手动打标签
<1s
File Submission Latency
文件提交响应时间
4
Relationship Types in Graph
知识图谱关系类型

From Inbox to Insight
in Three Steps
从收件箱到洞见,
只需三步

AiSyncMe eliminates the overhead of manual knowledge organization. Drop a file. Get a searchable, queryable knowledge entry — automatically.

AiSyncMe 彻底消除手动整理知识的负担。投入一个文件,自动获得结构化、可检索的知识条目。

01
📥

Drop Your File投入文件

Drag any file into your watched inbox folder — PDF, DOCX, MP3, screenshot, HTML page, or code file. AiSyncMe detects it within milliseconds and queues it for processing. No UI required.

将任何文件拖入受监控的收件箱文件夹——PDF、DOCX、MP3、截图、网页或代码文件皆可。AiSyncMe 毫秒级检测并自动排队处理,无需任何界面操作。

02
🤖

AI Reads & Structures ItAI 阅读并结构化

Gemini 2.5 extracts text, generates a summary, identifies entities (people, companies, concepts), assigns tags, and writes a structured Markdown knowledge entry — automatically.

Gemini 2.5 自动提取文本、生成摘要、识别实体(人物、公司、概念)、分配标签,并写入结构化 Markdown 知识条目——全程无需人工干预。

03
💬

Ask Anything随时提问

Use natural language to query your knowledge base. AiSyncMe finds the most relevant entries, synthesizes an answer using RAG, and cites its sources — like a personal research assistant.

用自然语言查询整个知识库。AiSyncMe 找出最相关的条目,通过 RAG 生成答案并标注来源——就像拥有一位私人研究助理。

Works with Every Format
You Already Use
兼容你日常使用的
所有文件格式

AiSyncMe reads files the way a human would — including image content in photos and speech in audio recordings.

AiSyncMe 像人类一样"读取"文件——包括图片中的视觉内容和音频中的语音内容。

📄 PDF
📝 Word / DOCX
📊 PowerPoint / PPTX
🖼️ Images (JPG, PNG, WebP)图片(JPG, PNG, WebP)
🎙️ Audio (MP3, M4A, WAV)音频(MP3, M4A, WAV)
🎬 Video (MP4)视频(MP4)
🌐 Web Pages / HTML网页 / HTML
📓 Markdown / NotesMarkdown / 笔记
💻 Code (Python, JS, TS…)代码(Python, JS, TS…)
📋 Plain Text / CSV纯文本 / CSV
+ 20 more formats+ 20 种更多格式

Everything Your Knowledge
Needs to Come Alive
让你的知识
真正活起来

Interactive Knowledge Graph交互式知识图谱

Your documents become an interconnected graph. Discover hidden connections between files through shared entities, tags, people, or explicit wiki-links — visualized in real time.

你的文档变成互联图谱。通过共享实体、标签、人物或双向链接,实时可视化发现文件之间隐藏的知识联系。

Force-directed layout力导向布局 4 relationship types4 种关系类型 Drag & pin nodes拖拽固定节点
💬

RAG Conversational AIRAG 对话式 AI

Ask multi-turn questions in plain language. The Hermes Agent retrieves the most semantically relevant entries and generates grounded answers with source citations for every claim.

用自然语言进行多轮对话。Hermes Agent 检索语义最相关的条目,生成有来源支撑的精准答案——每个结论均可溯源。

18-turn context18 轮上下文 Source citations来源引用 Streaming output流式输出
📥

Zero-Friction Inbox零摩擦收件箱

A watchdog service monitors any folder you designate. Files dropped in are automatically processed and entered into your knowledge base — no button clicks required.

watchdog 服务监控你指定的任意文件夹。投入的文件自动处理并入库——无需点击任何按钮。

Folder monitoring目录自动监控 Auto-processing自动处理 SHA-256 dedupSHA-256 去重
📊

AI Summarization & TaggingAI 摘要与自动标签

Every ingested file gets a structured Markdown entry with title, one-sentence summary, category, tags, and named entities — generated by Gemini 2.5, zero manual input required.

每个入库文件自动获得结构化 Markdown 条目:标题、一句话摘要、分类、标签、命名实体——由 Gemini 2.5 生成,全程零人工输入。

Gemini 2.5 Flash Entity extraction实体识别 Auto-categorization自动分类
🔍

Hybrid Search混合检索

Full-text FTS5 search combined with semantic vector retrieval and RRF fusion ranking. Keyword matching automatically kicks in as fallback when AI search is unavailable.

FTS5 全文检索结合语义向量检索,RRF 融合排序。AI 检索不可用时自动降级为关键词匹配,保证检索永不失效。

FTS5 Semantic vectors语义向量 RRF rankingRRF 融合排序
📅

Daily Knowledge Review每日知识复盘

Ask AiSyncMe to recap what you learned today or this week. It generates a structured review with new additions, cross-document insights, and actionable recommendations.

让 AiSyncMe 复盘今天或本周你学到了什么。它自动生成结构化复盘报告:新增内容摘要、跨文档知识连接洞见、行动建议。

Daily digest每日摘要 Weekly summary每周总结 Insight synthesis洞见生成

Ask Your Knowledge Base
Like You Would a Colleague
像问同事一样,
直接问你的知识库

AiSyncMe answers from your actual documents — not the internet. Every answer includes the source files it used, so you can verify instantly.

AiSyncMe 的答案来自你自己的文档——而非互联网。每条答案均附上所引用的原始文件,随时可验证。

Hermes Agent · Connected to Knowledge Base Hermes Agent · 已连接知识库 ⚡ Gemini 2.5 Flash
👤
What were the key action items from last week's product standup, and how do they relate to our Q2 strategy?
Based on your uploaded standup recording and Q2 strategy document, here are the key connections:

Action Items from Standup (May 12):
• Complete onboarding flow redesign by May 20 — aligns with Q2 Goal #2: Reduce time-to-value for new users
• Investigate API latency spike — flagged as risk in Q2 OKRs under Infrastructure Reliability
• Schedule user interview panel — supports Q2 Goal #3: Validate pricing tiers with 20 paying customers

The standup's focus on onboarding directly mirrors the Q2 priority of cutting activation drop-off from 62% to below 40%.
Team Standup May 12.m4a Q2 Strategy v3.pdf OKR Tracker Q2.docx
👤
上周产品站会有哪些关键行动项?它们和我们的 Q2 策略有什么关联?
根据你上传的站会录音和 Q2 策略文档,以下是关键关联:

站会行动项(5月12日):
• 5月20日前完成引导流程重设计 — 对应 Q2 目标 #2:缩短新用户从注册到获得价值的时间
• 排查 API 延迟异常 — 在 Q2 OKR 基础设施稳定性风险项中已标记
• 安排用户访谈 — 支撑 Q2 目标 #3:与 20 位付费用户验证定价方案

站会中对引导流程的关注,与 Q2 优先事项"将激活流失率从 62% 降至 40% 以下"完全吻合。
团队站会_5月12日.m4a Q2战略 v3.pdf Q2_OKR追踪表.docx

Common Questions 常见问题

What file formats does AiSyncMe support?
AiSyncMe 支持哪些文件格式?
AiSyncMe supports PDF, DOCX, PPTX, images (JPG, PNG, GIF, WebP), audio and video (MP3, MP4, M4A, WAV), HTML web pages, Markdown, and 30+ code and plain-text formats including Python, JavaScript, and TypeScript.
AiSyncMe 支持 PDF、DOCX、PPTX、图片(JPG/PNG/GIF/WebP)、音视频(MP3/MP4/M4A/WAV)、HTML 网页、Markdown,以及 30+ 种代码和纯文本格式,包括 Python、JavaScript、TypeScript 等。
How does AiSyncMe organize my knowledge automatically?
AiSyncMe 如何自动整理我的知识?
When you drop a file into your inbox, AiSyncMe extracts the text, then uses Gemini 2.5 to generate a structured summary including title, category, tags, entities, and key insights — indexed with FTS5 full-text search and semantic vector embeddings. Zero manual tagging.
当你将文件投入收件箱时,AiSyncMe 提取文本内容,然后用 Gemini 2.5 生成结构化摘要(标题、分类、标签、实体、关键洞见),并以 FTS5 全文索引和语义向量嵌入的方式入库。全程零手动标注。
Is my data stored locally or sent to the cloud?
我的数据是存在本地还是上传到云端?
Your knowledge entries are stored locally as Markdown files and SQLite on your own machine. AI processing uses the Gemini API, but your raw files stay on your device. You can read your notes in any Markdown editor like Obsidian.
你的知识条目以 Markdown 文件和 SQLite 数据库的形式存储在本地设备上。AI 处理调用 Gemini API,但原始文件始终在你的机器上。你可以用 Obsidian 等任何 Markdown 编辑器直接阅读笔记。
Can I use AiSyncMe without running it all the time?
AiSyncMe 需要一直运行吗?
Yes. The folder watchdog runs as a lightweight background service. The web UI and API can be started on demand. Everything is orchestrated with Docker Compose — one command to start, one to stop.
不需要。文件夹监控作为轻量后台服务运行,Web UI 和 API 服务可按需启动。所有服务通过 Docker Compose 编排——一条命令启动,一条命令停止。
What AI model powers AiSyncMe?
AiSyncMe 使用什么 AI 模型?
AiSyncMe uses Gemini 2.5 Flash for fast summarization and conversational Q&A, and Gemini 2.5 Pro for deep document and image analysis. Audio and video transcription uses Gemini's multimodal Audio API.
AiSyncMe 使用 Gemini 2.5 Flash 进行快速摘要生成和对话问答,使用 Gemini 2.5 Pro 进行复杂文档和图片深度分析,音视频转录使用 Gemini 多模态 Audio API。

Your Knowledge.
Finally Searchable.
你的知识,
终于可以被检索了。

Stop losing insights in a sea of files. Start building a knowledge brain that works for you. 不再让洞见淹没在文件的海洋里。开始构建一个真正为你工作的知识大脑。

Get Started Free免费开始使用 Star on GitHubGitHub 上 Star
Open Source · Self-Hosted · Your Data Stays Yours 开源 · 自托管 · 数据完全属于你