用 Python 写的小项目共 48 个，按热度、创意度和网友投票排序。同一个技术栈的项目放在一起看，能很快判断这个方向用什么做最省事。
Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per request.
Fastest and cheapest web agent
Native MLX runtime for Laya typed decision models — 7–14 ms short decisions on M3 Max. No text generation, PyTorch, or cloud API.
Jev-like family of decision models built on top of Qwen3.5/3.8 you can train and run on your own
Semantic ifs from open models, on a 3090 at home. Independent; not affiliated with Jev or TypeSafe.
Semantic ifs from open models, on a 3090 at home. Independent; not affiliated with Jev or TypeSafe.
Suitable for Android APK reverse engineering analysis
《深入理解 AI Infra：量化分析与系统设计》（李博杰 著）开源书稿：从硬件约束和模型架构出发，量化推导 LLM 推理与训练系统设计。含全书正文、PDF、配套计算工具与实验
A nano replica of Jev: parallel decisions, dynamic candidates, and an end-to-end training pipeline.
Local typed decisions, contrastive data curation, and model evaluation.
Private source preview: native Jianying drafts, isolated editing/export, and standalone Agent Skill.
Nosis 是一个开源的轻量级个人 Agent Runtime：执行边界由 Session 自行声明，越界命令在 OS 级沙箱中运行；内置工具集与 Plugin / Skill / MCP 扩展机制齐备，GUI与TUI共享Agent Runtime，会话、计划与记忆以文件持久化。
An experimental from-scratch image compression pipeline that turns pixels into text (grid averaging → RLE → alphanumeric mapping → LZW-style dictionary) and reconstructs the image back — no external compression libraries used.
Turn the user's description or uploaded reference into a finished, editable Blender card and an interactive Three.js page. Preserve the requested subject, style, typography and destination. This skill contains code and text only; generated artwork belongs in the user's output project.
Local-first WeChat intelligence system with a read-only CLI, Codex skills, searchable chat history, daily briefings, follow-ups and opportunity tracking.
A really simple and minimal (almost toy like, but not complete toy) library simulator written with minimal python dependencies
Qwen's most powerful open-source image generation model
De-AI writing skill for any Agent Skills-compatible agent (77+ via the Skills CLI), with native plugins for Claude Code, Codex, Grok Build, and Antigravity. Narrative-architecture repair for fiction, venue-matched rules for professional prose. Based on StoryScope (arXiv:2604.03136).
磁盘空间变化对比轻量小工具；A lightweight disk space change comparison tool.
Infrastructure for continually self‑improving agents
Autonomous research system for measurable, computer-executable research.
Use Astra or Sol as orchestrator and Luna for subagents in Codex
Reference blueprint for building shopping and merchant agents with Claude. Examples in retail, commerce, telecom, and entertainment included.
🧩 FrontierAgent, our agent framework, open-sourced alongside it — native command-line TUI, ReAct and Agent Team modes, one command on macOS and Linux, no preinstall, no hard Docker dependency.
One-ink editorial print image skill — warm paper, halftone photography, active negative space, and restrained typography.
一个基于 283 万字语料统计的去 AI 味 skill · An AI-tone removal skill grounded in a 2.83-million-character corpus study
WeMM-Embedding is a family of universal multimodal embedding models by the WeChat Vision Team at Tencent, supporting multimodal understanding and retrieval.
Amagine3D: From hardware requirements to editable 3D designs
从「龙猫体育锻炼」微信小程序自己已登录的进程里读出 token 的本地命令行工具
A privacy-first app that strips AI watermarks from content you own.
ARTEMIS turns natural-language instructions into reliable Android automation. It automates end-to-end workflows, captures logs, and integrates seamlessly with AI coding assistants such as Antigravity, Codex, and Claude Code. It also achieves 99%+ success rate on AndroidWorld Benchmark.
全平台加密音乐格式转换器（网页版+命令行）：网易云 NCM / QQ音乐 QMC / 酷狗 KGM / 酷我 KWM 解密与转码，代码从零自研，MIT 协议
小饼干英语视频工坊：从英语知识点到 30 秒竖屏教学视频的一站式 AI 创作工具。
小liu破甲 - Codex CLI jailbreak prompt and one-click deploy tool. 基于CTF框架的Codex破甲提示词与一键部署工具
Generates original ARC-AGI-1-style tasks distribution-matched to the public eval set.
Toy multi-tenant to-do app using Django and Vue
在终端（PowerShell / Windows Terminal / Linux Shell）中查看 A 股 K 线图、市盈率、成交量等基本信息的小工具。
知归是一个面向个人使用的 AI 知识归档工具。把内容链接（微信公众号、稀土掘金、GitHub、小红书、抖音、B站）发送给飞书机器人，知归会自动完成内容提取、AI 分类与总结，并将结果沉淀到飞书多维表格中。Send WeChat Article / Douyin / Xiaohongshu / Bilibili links to a Feishu (Lark) bot; auto-extract, AI-summarize and archive to Feishu Bitable. Self-hosted personal knowledge base / second brain.
Windows 小白向 LoRA 一键训练桌面工具：画风/人物/视频，支持 SD1.5/SDXL/FLUX/Anima/Krea2/H3/Qwen-Image/Z-Image，应用内下载 + 自动更新
多平台自媒体自动发布工具，支持抖音/B站/小红书/快手/视频号/百家号/头条/搜狐/知乎/TikTok，提供 Web、CLI 与示例脚本
台灣市場籌碼資料的 Python 小工具：期交所日行情與三大法人、證交所三大法人與融資融券
FreeToken brings datacenter-scale model serving to your desktop. Run massive models locally, fast and efficiently.
项目信息自动抓取自 GitHub 与 Hacker News，版权归原作者所有。完整榜单见 小项目灵感榜。