[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-73613":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":25,"hasPages":25,"topics":26,"createdAt":10,"pushedAt":10,"updatedAt":27,"readmeContent":28,"aiSummary":29,"trendingCount":16,"starSnapshotCount":16,"syncStatus":30,"lastSyncTime":31,"discoverSource":32},73613,"aipyapp","knownsec\u002Faipyapp","knownsec","AI-Powered Python & Python-Powered AI (Python-Use)","https:\u002F\u002Faipy.app",null,"HTML",3970,400,33,11,0,3,4,21,9,29.81,"Other",false,"main",true,[],"2026-06-12 02:03:15","![logo](https:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F3af4e228-79b2-4fa0-a45c-c38276c6db91)\n\n# Python-Use: A New AI Agent Paradigm (Agent 2.0)\n\n**AI-Powered Python & Python-Powered AI**\n\nPython-Use is a task-driven, result-oriented intelligent execution paradigm. It tightly integrates LLMs with a Python interpreter to establish a complete loop:\n\n> **Task → Plan → Code → Execute → Feedback**\n\n## Background: The Outdated \"Prosthetic\" AI Agent Model\n\nTraditional AI (Agent 1.0) relies on Function Calling, Tools, MCP-Servers, Workflows, and plugin-based clients. These external \"prosthetics\" lead to:\n- High entry barriers\n- Heavy reliance on developers\n- Poor coordination between tools\n- Most AI-generated code locked in cloud sandboxes, unable to interact with the real environment\n\nWe urgently need a new paradigm that reconnects AI with the real world and fully activates its native execution power—ushering in the **AI Think Do** era.\n\n## What is Python-Use?\n\nPython-Use provides the entire Python execution environment to LLM. Imagine LLM sitting in front of a computer, typing various commands into the Python command-line interpreter, pressing Enter to execute, observing the results, and then typing and executing more code.\n\nThis gives models two core capabilities:\n- **API Calling**: Automatically generate and execute Python code to invoke APIs\n- **Packages Calling**: Flexibly leverage Python's ecosystem to orchestrate workflows\n\nUsers only need to provide a task description or API key. The model handles the rest—no plugin registration, no toolchain setup, no workflow editing.\n\n> **Important**: Python-Use is _not_ a code generator or smart IDE.\n> It's a task-first, outcome-driven AI Agent.\n\nTo the user, Python-Use is simple:\n> Describe a task → AI executes it → Result returned.\n\nThe model autonomously understands, plans, writes, debugs, and executes code—and fixes bugs along the way. Code is just an internal implementation—not the deliverable. The real deliverable is the **result**.\n\n## Why Python?\n\nWhile this paradigm theoretically supports any language, we choose Python because:\n- It has a powerful ecosystem spanning data, automation, system control, and AI\n- Its syntax is simple and readable, ideal for model generation and debugging\n- Models are naturally more proficient in Python for accurate and efficient coding\n\n## Core Principle: No Agents, Code is Agent\n\nPython-Use introduces a radically simplified execution architecture:\n\n**No Agents, No MCP, No Workflow, No Clients…**\n\nIt discards legacy layers and lets models use code to directly act on the environment. In short: **Code is Agent**.\n\nWith Python, the model can:\n- **Python use Data**: Load, transform, analyze\n- **Python use Browser**: Automate the web\n- **Python use Computer**: Access file systems and local resources\n- **Python use IoT**: Control devices and embedded systems\n- **…**\n- **Python use Anything**: Code becomes a universal interface\n\nThis means:\n- **No MCP**: No standardized protocol needed—code is the protocol\n- **No Workflow**: Model plans and executes on the fly\n- **No Tools**: No plugin registrations needed—just use existing ecosystems\n- **No Agents**: Code replaces orchestration—execution becomes native\n\nThis is the bridge that reconnects LLMs to the real digital world, unlocking their latent power.\n\n## Execution Mode: AI Think Do\n\n**AI Think Do = True Integration of Knowing & Doing**\n\n- **Task**: User describes intent\n- **Plan**: Model decomposes and plans a path\n- **Code**: Optimal Python strategy is generated\n- **Execute**: Direct interaction with the environment\n- **Feedback**: Output is evaluated and looped back into planning\n\nNo external agent needed. The AI completes the full loop independently, unleashing true cognitive-action capability.\n\n## Single Entry Point: AiPy\n\nYou don't need multiple AI apps or UI wrappers anymore.\n\nJust run one thing: **AiPy**, a Python-powered AI Client.\n\n- **Unified interface**: All interaction via Python\n- **Zero clutter**: No plugin mess, no bloated clients\n- **AiPy**: https:\u002F\u002Fwww.aipy.app\u002F\n\n## Usage\nAiPy has two running modes:\n\n### Task Mode (Default)\nVery simple and easy to use—just input your task. Suitable for users unfamiliar with Python.\n\n### Python Mode (`--python`)\nSuitable for users familiar with Python, allowing both task input and Python commands. Ideal for advanced users.\n\n## Basic Config\n\nCreate `~\u002F.aipyapp\u002Faipyapp.toml`:\n\n```toml\n[llm.deepseek]\ntype = \"deepseek\"\napi_key = \"Your DeepSeek API Key\"\n```\n\n## Task Mode Examples\n\n### Installation\n```bash\npip install aipyapp\n```\n\n### Usage\n```bash\naipy\n```\n\n```\n🚀 Python use - AIPython (0.1.22) [https:\u002F\u002Faipy.app]\n>>> Get the latest posts from Reddit r\u002FLocalLLaMA\n......\n>>> \u002Fdone\n```\n\n## Python Mode Examples\n\n### Start Python Mode\n```bash\naipy --python\n```\n\n### Basic Usage\nAutomatic task processing:\n\n```\nPython use - AIPython (Quit with 'exit()')\n>>> ai(\"Get the title of Google's homepage\")\n```\n\n### Automatically Request to Install Third-Party Libraries\n```\n>>> ai(\"Use psutil to list all processes on MacOS\")\n\n📦 LLM requests to install third-party packages: ['psutil']\nIf you agree and have installed, please enter 'y [y\u002Fn] (n): y\n```\n\n### Basic Config\n~\u002F.aipyapp\u002Faipyapp.toml:\n```toml\n[llm.deepseek]\ntype = \"deepseek\"\napi_key = \"Your DeepSeek API Key\"\n```\n\n### Task Mode\n`uv run aipy`\n```\n>>> Get the latest posts from Reddit r\u002FLocalLLaMA\n......\n......\n>>> \u002Fdone\n```\n\n`pip install aipyapp` and run with `aipy`\n\n```\n-> % aipy\n🚀 Python use - AIPython (0.1.22) [https:\u002F\u002Faipy.app]\n>> Get the latest posts from Reddit r\u002FLocalLLaMA\n......\n>>\n```\n\n### Python Mode\n\n#### Basic Usage\nAutomatic task processing:\n\n```\n>>> ai(\"Get the title of Google's homepage\")\n```\n\n#### Automatically Request to Install Third-Party Libraries\n```\nPython use - AIPython (Quit with 'exit()')\n>>> ai(\"Use psutil to list all processes on MacOS\")\n\n📦 LLM requests to install third-party packages: ['psutil']\nIf you agree and have installed, please enter 'y [y\u002Fn] (n): y\n\n```\n\n## Vision: Free the AI, Reach AGI\n\nPython-Use is more than a tool—it's a future-facing AI philosophy:\n\n> **The Model is the Product → The Model is the Agent → No Agents, Code is Agent → Just Python-use → Freedom AI (AGI)**\n\nIt transforms AI from \"just speaking\" to \"taking action,\" from plugin-bound to autonomous execution. It unlocks full production power—and lights the path to general intelligence.\n\nJoin us. Let AI break free, act freely, and build the future.\n\n**The real general AI Agent is NO Agents!**\n\n**No Agents, Just Python-use!**\n\n## Self-Evolution: Multi-Model Fusion\n\nAI evolution is not just language modeling—it's multi-modal intelligence.\n\n- Integrates vision models for image\u002Fvideo understanding\n- Adds speech models for listening and speaking\n- Embeds expert models for domain reasoning\n- All fused and coordinated under a unified AI control loop\n\nThis moves us from \"chatbots\" to fully embodied AI agents—on the path to true AGI.\n\n## Thanks\n\n- **Hei Ge**: Product manager\u002Fsenior user\u002Fchief tester\n- **Sonnet 3.7**: Generated the first version of the code, which was almost ready to use without modification\n- **ChatGPT**: Provided many suggestions and code snippets, especially for the command-line interface\n- **Codeium**: Intelligent code completion\n- **Copilot**: Code improvement suggestions\n\n---\n\n**Python-Use: The Future of AI Agents**\n\n\n\n","knownsec\u002Faipyapp 是一个基于AI的任务驱动、结果导向的智能执行框架，它将大语言模型与Python解释器紧密结合，形成任务规划、代码生成、执行及反馈的完整闭环。其核心功能包括自动API调用和灵活利用Python生态系统来编排工作流，使得用户只需提供任务描述或API密钥即可完成复杂任务，无需手动配置工具链或编辑工作流。该项目特别适合需要高效处理数据、自动化操作、系统控制等场景，尤其是当任务涉及多步骤逻辑且希望减少开发人员直接介入时。通过采用简洁而强大的Python语言作为实现手段，aipyapp 能够在保持易读性和可维护性的同时，充分发挥AI的自主执行能力。",2,"2026-06-11 03:46:24","high_star"]