[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-72620":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":23,"hasPages":23,"topics":25,"createdAt":10,"pushedAt":10,"updatedAt":32,"readmeContent":33,"aiSummary":34,"trendingCount":16,"starSnapshotCount":16,"syncStatus":35,"lastSyncTime":36,"discoverSource":37},72620,"open-computer-use","e2b-dev\u002Fopen-computer-use","e2b-dev","AI computer use powered by open source LLMs and E2B Desktop Sandbox","https:\u002F\u002Fe2b.dev",null,"Python",2067,242,21,10,0,14,24,64,42,97.56,"Apache License 2.0",false,"master",[26,27,28,29,30,31],"agent","ai","anthropic","claude","computer-use","llm","2026-06-12 04:01:06","# Open Computer Use\n\nA secure cloud Linux computer powered by [E2B Desktop Sandbox](https:\u002F\u002Fgithub.com\u002Fe2b-dev\u002Fdesktop\u002F) and controlled by open-source LLMs.\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F3837c4f6-45cb-43f2-9d51-a45f742424d4\n\n## Features\n\n- Uses [E2B](https:\u002F\u002Fe2b.dev) for secure [Desktop Sandbox](https:\u002F\u002Fgithub.com\u002Fe2b-dev\u002Fdesktop)\n- Operates the computer via the keyboard, mouse, and shell commands\n- Supports 10+ LLMs, [OS-Atlas](https:\u002F\u002Fosatlas.github.io\u002F)\u002F[ShowUI](https:\u002F\u002Fgithub.com\u002Fshowlab\u002FShowUI) and [any other models you want to integrate](#llm-support)!\n- Live streams the display of the sandbox on the client computer\n- User can pause and prompt the agent at any time\n- Uses Ubuntu, but designed to work with any operating system\n\n## Design\n\n![Open Computer Use Architecture](.\u002Fassets\u002Farchitecture.png#gh-dark-mode-only)\n![Open Computer Use Architecture](.\u002Fassets\u002Farchitecture-light.png#gh-light-mode-only)\n\nThe details of the design are laid out in this article: [How I taught an AI to use a computer](https:\u002F\u002Fblog.jamesmurdza.com\u002Fhow-i-taught-an-ai-to-use-a-computer)\n\n## LLM support\n\nOpen Computer Use is designed to make it easy to swap in and out new LLMs. The LLMs used by the agent are specified in [config.py](\u002Fos_computer_use\u002Fconfig.py) like this:\n\n```\ngrounding_model = providers.OSAtlasProvider()\nvision_model = providers.GroqProvider(\"llama3.2\")\naction_model = providers.GroqProvider(\"llama3.3\")\n```\n\nThe providers are imported from [providers.py](\u002Fos_computer_use\u002Fproviders.py) and include:\n\n- Fireworks, OpenRouter, Llama API:\n  - Llama 3.2 (vision only), Llama 3.3 (action only)\n- Groq:\n  - Llama 3.2 (vision + action), Llama 3.3 (action only)\n- DeepSeek:\n  - DeepSeek (action only)\n- Google:\n  - Gemini 2.0 Flash (vision + action)\n- OpenAI:\n  - GPT-4o and GPT-4o mini (vision + action)\n- Anthropic:\n  - Claude (vision + action)\n- HuggingFace Spaces:\n  - OS-Atlas (grounding)\n  - ShowUI (grounding)\n- Moonshot\n- Mistral AI (Pixtral for vision, Mistral Large for actions)\n\nIf you add a new model or provider, please [make a PR](..\u002F..\u002Fpulls) to this repository with the updated providers.py!\n\n## Get started\n\n### Prerequisites\n\n- Python 3.10 or later\n- [git](https:\u002F\u002Fgit-scm.com\u002F)\n- [E2B API key](https:\u002F\u002Fe2b.dev\u002Fdashboard?tab=keys)\n- API key for an LLM provider (see above)\n\n### 1. Install the prerequisites\n\nIn your terminal:\n\n```sh\nbrew install poetry ffmpeg\n```\n\n### 2. Clone the repository\n\nIn your terminal:\n\n```sh\ngit clone https:\u002F\u002Fgithub.com\u002Fe2b-dev\u002Fopen-computer-use\u002F\n```\n\n### 3. Set the environment variables\n\nEnter the project directory:\n\n```\ncd open-computer-use\n```\n\nCreate a `.env` file in `open-computer-use` and set the following:\n\n```sh\n# Get your API key here: https:\u002F\u002Fe2b.dev\u002F\nE2B_API_KEY=\"your-e2b-api-key\"\n```\n\nAdditionally, add API key(s) for any LLM providers you're using:\n```\n# You only need the API key for the provider(s) selected in config.py:\n# Hugging Face Spaces do not require an API key.\nFIREWORKS_API_KEY=...\nOPENROUTER_API_KEY=...\nLLAMA_API_KEY=...\nGROQ_API_KEY=...\nGEMINI_API_KEY=...\nOPENAI_API_KEY=...\nANTHROPIC_API_KEY=...\nMOONSHOT_API_KEY=...\n# Required: Provide your Hugging Face token to bypass Gradio rate limits.\nHF_TOKEN=...\n```\n\n### 4. Start the web interface\n\nRun the following command to start the agent:\n\n```sh\npoetry install\n```\n\n```sh\npoetry run start\n```\n\nThe agent will open and prompt you for its first instruction.\n\nTo start the agent with a specified prompt, run:\n\n```sh\npoetry run start --prompt \"use the web browser to get the current weather in sf\"\n```\n\nThe display stream should be visible a few seconds after the Python program starts.\n\n","Open Computer Use 是一个基于开源大语言模型和E2B桌面沙箱的AI控制云Linux计算机项目。该项目通过模拟键盘、鼠标操作及执行shell命令来操控虚拟计算机，并支持超过10种大语言模型，包括OS-Atlas、ShowUI等，且设计上允许用户轻松更换所使用的LLM。它还提供了实时显示沙箱内活动的功能，用户可以随时暂停并指导AI代理的行为。此工具非常适合需要在安全可控环境下进行软件测试、开发教学或研究AI与操作系统交互机制的应用场景。",2,"2026-06-11 03:42:49","high_star"]