## 内容
> **以下为本期内容**|每月 **28** 号更新
### C 项目
1、[iotop](https://hellogithub.com/periodical/statistics/click?target=https://github.com/Tomas-M/iotop):终端下的 IO 资源监控工具。这是一款用于监控 Linux 系统 I/O 的命令行工具,拥有类似 top 命令的交互界面和操作方式,支持实时按 I/O 使用率对进程进行排序和展示。
2、[libsodium](https://hellogithub.com/periodical/statistics/click?target=https://github.com/jedisct1/libsodium):开箱即用的 C 语言加密库。这是一个现代易用、跨平台的 C 语言加密库,为开发者提供全面的加密操作 API。它集成了多种加密、签名和哈希算法,适用于安全通信、数据保护等场景。来自 [@沐怡旸](https://hellogithub.com/user/76dhrs54nQGLuge) 的分享
27、[yaak](https://hellogithub.com/periodical/statistics/click?target=https://github.com/mountain-loop/yaak):离线优先的桌面 API 客户端。这是一款快速、离线优先的桌面 API 客户端,支持 REST、GraphQL、SSE、WebSocket 和 gRPC 协议。它基于 Rust、Tauri 和 React 构建,界面友好、跨平台可用,无遥测和云端锁定,提供简洁纯粹、不受干扰的使用体验。来自 [@刘睿华](https://hellogithub.com/user/TJ65FfbQU09PLHM) 的分享
### Swift 项目
28、[bitchat](https://hellogithub.com/periodical/statistics/click?target=https://github.com/permissionlesstech/bitchat):基于蓝牙的即时通信应用。这是一款无需服务器的蓝牙即时通讯应用,专为无网络环境设计。它通过结合蓝牙 Mesh 和 Nostr 协议,实现端到端加密的点对点通信,支持消息重试、离线转发、多用户、自动发现等功能,适用于救灾志愿者、野外探险队等小范围即时沟通场景。
29、[Dayflow](https://hellogithub.com/periodical/statistics/click?target=https://github.com/JerryZLiu/Dayflow):自动生成每日时间线的 macOS 应用。这是一款用 Swift 开发的 macOS 应用,通过录制屏幕活动并结合 AI,自动生成每日时间线。它以每秒 1 帧的频率录制屏幕,每 15 分钟利用 AI 分析录像内容,生成简洁的活动总结。同时,自动删除超过 3 天的录制文件,节省存储空间。
### 人工智能
30、[BrowserOS](https://hellogithub.com/periodical/statistics/click?target=https://github.com/browseros-ai/BrowserOS):开源的 AI 浏览器。该项目是基于 Chromium 的开源 AI 浏览器,能够在本地浏览器中运行 AI Agents,可作为 ChatGPT Atlas、Perplexity Comet 和 Dia 的开源替代方案。在保留 Chrome 熟悉界面与扩展兼容性的同时,帮助用户实现 AI 驱动的浏览器自动化与智能问答任务,并支持自定义 LLM 服务或本地大模型。
31、[cache-dit](https://hellogithub.com/periodical/statistics/click?target=https://github.com/vipshop/cache-dit):免训练的 DiT 模型缓存加速框架。该项目是为 Diffusers 提供统一缓存加速的框架,支持几乎所有的 DiT 扩散模型,包括 Qwen-Image-Lightning、Qwen-Image、HunyuanImage、Wan、FLUX 等。它通过简单的代码即可实现高效的缓存加速功能,无需重新训练模型即可显著提升推理速度。来自 [@DefTruth](https://hellogithub.com/user/ofSCbzTmdeQk3FD) 的分享
```python
import cache_dit
from diffusers import DiffusionPipeline
pipe = DiffusionPipeline.from_pretrained("Qwen/Qwen-Image") # Can be any diffusion pipeline
cache_dit.enable_cache(pipe) # One-line code with default cache options.
output = pipe(...) # Just call the pipe as normal.
stats = cache_dit.summary(pipe) # Then, get the summary of cache acceleration stats.
cache_dit.disable_cache(pipe) # Disable cache and run original pipe.
```
32、[graphiti](https://hellogithub.com/periodical/statistics/click?target=https://github.com/getzep/graphiti):为 AI 智能体构建实时知识图谱。这是一个专为 AI 智能体设计的框架,用于构建和查询实时、具有时间感知能力的知识图谱。它能够持续集成用户交互、结构化或非结构化等动态数据,形成连贯且可查询的知识图谱。支持增量数据更新、高效检索和历史查询,适用于开发交互式、上下文感知的 AI 应用。来自 [@塔咖](https://hellogithub.com/user/bzJpGyu0IanC6L7) 的分享
33、[LEANN](https://hellogithub.com/periodical/statistics/click?target=https://github.com/StarTrail-org/LEANN):超低存储占用的向量数据库。这是一款开源的轻量级向量数据库,通过按需计算嵌入向量,实现极低的存储占用。用户可以在个人设备(笔记本电脑)上构建强大且完全私有的检索增强生成 (RAG) 系统,支持对本地文件、电子邮件、浏览器历史、聊天记录等多种数据源进行语义搜索。来自 [@Yichuan Wang](https://hellogithub.com/user/Tj5AEF9BDNXpKnk) 的分享
```python
from leann import LeannBuilder, LeannSearcher, LeannChat
from pathlib import Path
INDEX_PATH = str(Path("./").resolve() / "demo.leann")
# Build an index
builder = LeannBuilder(backend_name="hnsw")
builder.add_text("LEANN saves 97% storage compared to traditional vector databases.")
builder.add_text("Tung Tung Tung Sahur called—they need their banana‑crocodile hybrid back")
builder.build_index(INDEX_PATH)
# Search
searcher = LeannSearcher(INDEX_PATH)
results = searcher.search("fantastical AI-generated creatures", top_k=1)
# Chat with your data
chat = LeannChat(INDEX_PATH, llm_config={"type": "hf", "model": "Qwen/Qwen3-0.6B"})
response = chat.ask("How much storage does LEANN save?", top_k=1)
```
34、[surf](https://hellogithub.com/periodical/statistics/click?target=https://github.com/deta/surf):跨媒体的个人 AI 笔记应用。这是一款本地优先的 AI 笔记本工具,能够将多种媒体类型(如本地文件、网页、视频等)整合至本地资料库,并借助 AI 快速生成笔记。它帮助用户在学习和研究过程中,免去在浏览器、笔记应用、PDF 阅读器等多个应用和媒体之间切换、搜索和手动复制粘贴的繁琐操作,同时支持灵活选择 AI 模型。
### 其它
35、[anki-jlpt-decks](https://hellogithub.com/periodical/statistics/click?target=https://github.com/5mdld/anki-jlpt-decks):万词语音全配的日语卡组。这是一个为学习软件 Anki 制作的高质量日语单词卡组,涵盖 JLPT(日本语能力测试)N1~N5 等级,共计一万词汇,每个词条都包含释义、例句、词性,以及关联词和反义词等。
36、[berkeley-humanoid-lite](https://hellogithub.com/periodical/statistics/click?target=https://github.com/HybridRobotics/berkeley-humanoid-lite):开源的人形机器人。该项目是由伯克利 Hybrid Robotics 团队开源,旨在提供低成本、模块化、可定制的人形机器人解决方案。机器人采用 3D 打印和常见组件,整体成本在 5000 美元以内,适用于机器人科研、算法开发和教学实验等多个领域。
39、[NCE](https://hellogithub.com/periodical/statistics/click?target=https://github.com/iChochy/NCE):《新概念英语》全四册在线课文朗读。该项目为《新概念英语》学习者提供了一个方便的在线学习平台,结合美式发音音频与 Gemini AI 生成的中文字幕,支持课文朗读与单句点读。