[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-4650":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":23,"topics":26,"createdAt":10,"pushedAt":10,"updatedAt":27,"readmeContent":28,"aiSummary":29,"trendingCount":16,"starSnapshotCount":16,"syncStatus":30,"lastSyncTime":31,"discoverSource":32},4650,"picoclaw","sipeed\u002Fpicoclaw","sipeed","Tiny, Fast, and Deployable anywhere — automate the mundane, unleash your creativity","https:\u002F\u002Fpicoclaw.io",null,"Go",29364,4207,142,42,0,15,83,429,77,120,"MIT License",false,"main",true,[],"2026-06-12 04:00:22","\u003Cdiv align=\"center\">\n\u003Cimg src=\"assets\u002Flogo.webp\" alt=\"PicoClaw\" width=\"512\">\n\n\u003Ch1>PicoClaw: Ultra-Efficient AI Assistant in Go\u003C\u002Fh1>\n\n\u003Ch3>$10 Hardware · 10MB RAM · ms Boot · Let's Go, PicoClaw!\u003C\u002Fh3>\n  \u003Cp>\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FGo-1.25+-00ADD8?style=flat&logo=go&logoColor=white\" alt=\"Go\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FArch-x86__64%2C%20ARM64%2C%20MIPS%2C%20RISC--V%2C%20LoongArch-blue\" alt=\"Hardware\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-green\" alt=\"License\">\n    \u003Cbr>\n    \u003Ca href=\"https:\u002F\u002Fpicoclaw.io\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FWebsite-picoclaw.io-blue?style=flat&logo=google-chrome&logoColor=white\" alt=\"Website\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.picoclaw.io\u002F\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDocs-Official-007acc?style=flat&logo=read-the-docs&logoColor=white\" alt=\"Docs\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdeepwiki.com\u002Fsipeed\u002Fpicoclaw\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FWiki-DeepWiki-FFA500?style=flat&logo=wikipedia&logoColor=white\" alt=\"Wiki\">\u003C\u002Fa>\n    \u003Cbr>\n    \u003Ca href=\"https:\u002F\u002Fx.com\u002FSipeedIO\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FX_(Twitter)-SipeedIO-black?style=flat&logo=x&logoColor=white\" alt=\"Twitter\">\u003C\u002Fa>\n    \u003Ca href=\".\u002Fassets\u002Fwechat.png\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FWeChat-Group-41d56b?style=flat&logo=wechat&logoColor=white\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FV4sAZ9XWpN\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDiscord-Community-4c60eb?style=flat&logo=discord&logoColor=white\" alt=\"Discord\">\u003C\u002Fa>\n  \u003C\u002Fp>\n\n[中文](docs\u002Fproject\u002FREADME.zh.md) | [日本語](docs\u002Fproject\u002FREADME.ja.md) | [한국어](docs\u002Fproject\u002FREADME.ko.md) | [Português](docs\u002Fproject\u002FREADME.pt-br.md) | [Tiếng Việt](docs\u002Fproject\u002FREADME.vi.md) | [Français](docs\u002Fproject\u002FREADME.fr.md) | [Italiano](docs\u002Fproject\u002FREADME.it.md) | [Bahasa Indonesia](docs\u002Fproject\u002FREADME.id.md) | [Malay](docs\u002Fproject\u002FREADME.ms.md) | **English**\n\n\u003C\u002Fdiv>\n\n---\n\n> **PicoClaw** is an independent open-source project initiated by [Sipeed](https:\u002F\u002Fsipeed.com), written entirely in **Go** from scratch — not a fork of OpenClaw, NanoBot, or any other project.\n\n**PicoClaw** is an ultra-lightweight personal AI assistant inspired by [NanoBot](https:\u002F\u002Fgithub.com\u002FHKUDS\u002Fnanobot). It was rebuilt from the ground up in **Go** through a \"self-bootstrapping\" process — the AI Agent itself drove the architecture migration and code optimization.\n\n**Runs on $10 hardware with \u003C10MB RAM** — that's 99% less memory than OpenClaw and 98% cheaper than a Mac mini!\n\n\u003Ctable align=\"center\">\n\u003Ctr align=\"center\">\n\u003Ctd align=\"center\" valign=\"top\">\n\u003Cp align=\"center\">\n\u003Cimg src=\"assets\u002Fpicoclaw_mem.gif\" width=\"360\" height=\"240\">\n\u003C\u002Fp>\n\u003C\u002Ftd>\n\u003Ctd align=\"center\" valign=\"top\">\n\u003Cp align=\"center\">\n\u003Cimg src=\"assets\u002Flicheervnano.png\" width=\"400\" height=\"240\">\n\u003C\u002Fp>\n\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftable>\n\n> [!CAUTION]\n> **Security Notice**\n>\n> * **NO CRYPTO:** PicoClaw has **not** issued any official tokens or cryptocurrency. All claims on `pump.fun` or other trading platforms are **scams**.\n> * **OFFICIAL DOMAIN:** The **ONLY** official website is **[picoclaw.io](https:\u002F\u002Fpicoclaw.io)**, and company website is **[sipeed.com](https:\u002F\u002Fsipeed.com)**\n> * **BEWARE:** Many `.ai\u002F.org\u002F.com\u002F.net\u002F...` domains have been registered by third parties. Do not trust them.\n> * **NOTE:** PicoClaw is in early rapid development. There may be unresolved security issues. Do not deploy to production before v1.0.\n> * **NOTE:** PicoClaw has recently merged many PRs. Recent builds may use 10-20MB RAM. Resource optimization is planned after feature stabilization.\n\n## 📢 News\n\n2026-03-31 📱 **Android Support!** PicoClaw now runs on Android! Download the APK at [picoclaw.io](https:\u002F\u002Fpicoclaw.io\u002Fdownload)\n\n2026-03-25 🚀 **v0.2.4 Released!** Agent architecture overhaul (SubTurn, Hooks, Steering, EventBus), WeChat\u002FWeCom integration, security hardening (.security.yml, sensitive data filtering), new providers (AWS Bedrock, Azure, Xiaomi MiMo), and 35 bug fixes. PicoClaw has reached **26K Stars**!\n\n2026-03-17 🚀 **v0.2.3 Released!** System tray UI (Windows & Linux), sub-agent status query (`spawn_status`), experimental Gateway hot-reload, Cron security gating, and 2 security fixes. PicoClaw has reached **25K Stars**!\n\n2026-03-09 🎉 **v0.2.1 — Biggest update yet!** MCP protocol support, 4 new channels (Matrix\u002FIRC\u002FWeCom\u002FDiscord Proxy), 3 new providers (Kimi\u002FMinimax\u002FAvian), vision pipeline, JSONL memory store, model routing.\n\n2026-02-28 📦 **v0.2.0** released with Docker Compose and Web UI Launcher support.\n\n\u003Cdetails>\n\u003Csummary>Earlier news...\u003C\u002Fsummary>\n\n2026-02-26 🎉 PicoClaw hits **20K Stars** in just 17 days! Channel auto-orchestration and capability interfaces are live.\n\n2026-02-16 🎉 PicoClaw breaks 12K Stars in one week! Community maintainer roles and [Roadmap](ROADMAP.md) officially launched.\n\n2026-02-13 🎉 PicoClaw breaks 5000 Stars in 4 days! Project roadmap and developer groups in progress.\n\n2026-02-09 🎉 **PicoClaw Released!** Built in 1 day to bring AI Agents to $10 hardware with \u003C10MB RAM. Let's Go, PicoClaw!\n\n\u003C\u002Fdetails>\n\n## ✨ Features\n\n🪶 **Ultra-lightweight**: Core memory footprint \u003C10MB — 99% smaller than OpenClaw.*\n\n💰 **Minimal cost**: Efficient enough to run on $10 hardware — 98% cheaper than a Mac mini.\n\n⚡️ **Lightning-fast boot**: 400x faster startup. Boots in \u003C1s even on a 0.6GHz single-core processor.\n\n🌍 **Truly portable**: Single binary across RISC-V, ARM, MIPS, and x86 architectures. One binary, runs everywhere!\n\n🤖 **AI-bootstrapped**: Pure Go native implementation — 95% of core code was generated by an Agent and fine-tuned through human-in-the-loop review.\n\n🔌 **MCP support**: Native [Model Context Protocol](https:\u002F\u002Fmodelcontextprotocol.io\u002F) integration — connect any MCP server to extend Agent capabilities.\n\n👁️ **Vision pipeline**: Send images and files directly to the Agent — automatic base64 encoding for multimodal LLMs.\n\n🧠 **Smart routing**: Rule-based model routing — simple queries go to lightweight models, saving API costs.\n\n_*Recent builds may use 10-20MB due to rapid PR merges. Resource optimization is planned. Boot speed comparison based on 0.8GHz single-core benchmarks (see table below)._\n\n\u003Cdiv align=\"center\">\n\n|                                | OpenClaw      | NanoBot                  | **PicoClaw**                           |\n| ------------------------------ | ------------- | ------------------------ | -------------------------------------- |\n| **Language**                   | TypeScript    | Python                   | **Go**                                 |\n| **RAM**                        | >1GB          | >100MB                   | **\u003C 10MB***                            |\n| **Boot time**\u003C\u002Fbr>(0.8GHz core) | >500s         | >30s                     | **\u003C1s**                                |\n| **Cost**                       | Mac Mini $599 | Most Linux boards ~$50   | **Any Linux board**\u003C\u002Fbr>**from $10**   |\n\n\u003Cimg src=\"assets\u002Fcompare.jpg\" alt=\"PicoClaw\" width=\"512\">\n\n\u003C\u002Fdiv>\n\n> **[Hardware Compatibility List](docs\u002Fguides\u002Fhardware-compatibility.md)** — See all tested boards, from $5 RISC-V to Raspberry Pi to Android phones. Your board not listed? Submit a PR!\n\n\u003Cp align=\"center\">\n\u003Cimg src=\"assets\u002Fhardware-banner.jpg\" alt=\"PicoClaw Hardware Compatibility\" width=\"100%\">\n\u003C\u002Fp>\n\n## 🦾 Demonstration\n\n### 🛠️ Standard Assistant Workflows\n\n\u003Ctable align=\"center\">\n\u003Ctr align=\"center\">\n\u003Cth>\u003Cp align=\"center\">Full-Stack Engineer Mode\u003C\u002Fp>\u003C\u002Fth>\n\u003Cth>\u003Cp align=\"center\">Logging & Planning\u003C\u002Fp>\u003C\u002Fth>\n\u003Cth>\u003Cp align=\"center\">Web Search & Learning\u003C\u002Fp>\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd align=\"center\">\u003Cp align=\"center\">\u003Cimg src=\"assets\u002Fpicoclaw_code.gif\" width=\"240\" height=\"180\">\u003C\u002Fp>\u003C\u002Ftd>\n\u003Ctd align=\"center\">\u003Cp align=\"center\">\u003Cimg src=\"assets\u002Fpicoclaw_memory.gif\" width=\"240\" height=\"180\">\u003C\u002Fp>\u003C\u002Ftd>\n\u003Ctd align=\"center\">\u003Cp align=\"center\">\u003Cimg src=\"assets\u002Fpicoclaw_search.gif\" width=\"240\" height=\"180\">\u003C\u002Fp>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd align=\"center\">Develop · Deploy · Scale\u003C\u002Ftd>\n\u003Ctd align=\"center\">Schedule · Automate · Remember\u003C\u002Ftd>\n\u003Ctd align=\"center\">Discover · Insights · Trends\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftable>\n\n### 🐜 Innovative Low-Footprint Deployment\n\nPicoClaw can be deployed on virtually any Linux device!\n\n- $9.9 [LicheeRV-Nano](https:\u002F\u002Fwww.aliexpress.com\u002Fitem\u002F1005006519668532.html) E(Ethernet) or W(WiFi6) edition, for a minimal home assistant\n- $30~50 [NanoKVM](https:\u002F\u002Fwww.aliexpress.com\u002Fitem\u002F1005007369816019.html), or $100 [NanoKVM-Pro](https:\u002F\u002Fwww.aliexpress.com\u002Fitem\u002F1005010048471263.html), for automated server operations\n- $50 [MaixCAM](https:\u002F\u002Fwww.aliexpress.com\u002Fitem\u002F1005008053333693.html) or $100 [MaixCAM2](https:\u002F\u002Fwww.kickstarter.com\u002Fprojects\u002Fzepan\u002Fmaixcam2-build-your-next-gen-4k-ai-camera), for smart surveillance\n\n\u003Chttps:\u002F\u002Fprivate-user-images.githubusercontent.com\u002F83055338\u002F547056448-e7b031ff-d6f5-4468-bcca-5726b6fecb5c.mp4>\n\n🌟 More Deployment Cases Await!\n\n## 📦 Install\n\n### Download from picoclaw.io (Recommended)\n\nVisit **[picoclaw.io](https:\u002F\u002Fpicoclaw.io)** — the official website auto-detects your platform and provides one-click download. No need to manually pick an architecture.\n\n### Download precompiled binary\n\nAlternatively, download the binary for your platform from the [GitHub Releases](https:\u002F\u002Fgithub.com\u002Fsipeed\u002Fpicoclaw\u002Freleases) page.\n\n### Build from source (for development)\n\nPrerequisites:\n\n- Go 1.25+\n- Node.js 22+ and pnpm 10.33.0+ for Web UI \u002F launcher builds\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fsipeed\u002Fpicoclaw.git\n\ncd picoclaw\nmake deps\n\n# Install frontend dependencies\n(cd web\u002Ffrontend && pnpm install --frozen-lockfile)\n\n# Build the core binary for the current platform\nmake build\n\n# Build the Web UI Launcher (required for WebUI mode)\nmake build-launcher\n\n# Build core binaries for all Makefile-managed platforms\nmake build-all\n\n# Build for Raspberry Pi Zero 2 W\n# 32-bit: make build-linux-arm\n# 64-bit: make build-linux-arm64\nmake build-pi-zero\n\n# Build and install\nmake install\n```\n\n**Raspberry Pi Zero 2 W:** Use the binary that matches your OS: 32-bit Raspberry Pi OS -> `make build-linux-arm`; 64-bit -> `make build-linux-arm64`. Or run `make build-pi-zero` to build both.\n\n## 🚀 Quick Start Guide\n\n### 🌐 WebUI Launcher (Recommended for Desktop)\n\nThe WebUI Launcher provides a browser-based interface for configuration and chat. This is the easiest way to get started — no command-line knowledge required.\n\n**Option 1: Double-click (Desktop)**\n\nAfter downloading from [picoclaw.io](https:\u002F\u002Fpicoclaw.io), double-click `picoclaw-launcher` (or `picoclaw-launcher.exe` on Windows). Your browser will open automatically at `http:\u002F\u002Flocalhost:18800`.\n\n**Option 2: Command line**\n\n```bash\npicoclaw-launcher\n# Open http:\u002F\u002Flocalhost:18800 in your browser\n```\n\n> [!TIP]\n> **Remote access \u002F Docker \u002F VM:** Add the `-public` flag to listen on all interfaces:\n> ```bash\n> picoclaw-launcher -public\n> ```\n\n\u003Cp align=\"center\">\n\u003Cimg src=\"assets\u002Flauncher-webui.jpg\" alt=\"WebUI Launcher\" width=\"600\">\n\u003C\u002Fp>\n\n**Getting started:**\n\nOpen the WebUI, then: **1)** Configure a Provider (add your LLM API key) -> **2)** Configure a Channel (e.g., Telegram) -> **3)** Start the Gateway -> **4)** Chat!\n\nFor detailed WebUI documentation, see [docs.picoclaw.io](https:\u002F\u002Fdocs.picoclaw.io).\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>Docker (alternative)\u003C\u002Fb>\u003C\u002Fsummary>\n\n```bash\n# 1. Clone this repo\ngit clone https:\u002F\u002Fgithub.com\u002Fsipeed\u002Fpicoclaw.git\ncd picoclaw\n\n# 2. First run — auto-generates docker\u002Fdata\u002Fconfig.json then exits\n#    (only triggers when both config.json and workspace\u002F are missing)\ndocker compose -f docker\u002Fdocker-compose.yml --profile launcher up\n# The container prints \"First-run setup complete.\" and stops.\n\n# 3. Set your API keys\nvim docker\u002Fdata\u002Fconfig.json\n\n# 4. Start\ndocker compose -f docker\u002Fdocker-compose.yml --profile launcher up -d\n# Open http:\u002F\u002Flocalhost:18800\n```\n\n> **Docker \u002F VM users:** The Gateway listens on `127.0.0.1` by default. Set `PICOCLAW_GATEWAY_HOST=0.0.0.0` or use the `-public` flag to make it accessible from the host.\n\n```bash\n# Check logs\ndocker compose -f docker\u002Fdocker-compose.yml logs -f\n\n# Stop\ndocker compose -f docker\u002Fdocker-compose.yml --profile launcher down\n\n# Update\ndocker compose -f docker\u002Fdocker-compose.yml pull\ndocker compose -f docker\u002Fdocker-compose.yml --profile launcher up -d\n```\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>macOS — First Launch Security Warning\u003C\u002Fb>\u003C\u002Fsummary>\n\nmacOS may block `picoclaw-launcher` on first launch because it is downloaded from the internet and not notarized through the Mac App Store.\n\n**Step 1:** Double-click `picoclaw-launcher`. You will see a security warning:\n\n\u003Cp align=\"center\">\n\u003Cimg src=\"assets\u002Fmacos-gatekeeper-warning.jpg\" alt=\"macOS Gatekeeper warning\" width=\"400\">\n\u003C\u002Fp>\n\n> *\"picoclaw-launcher\" Not Opened — Apple could not verify \"picoclaw-launcher\" is free of malware that may harm your Mac or compromise your privacy.*\n\n**Step 2:** Open **System Settings** → **Privacy & Security** → scroll down to the **Security** section → click **Open Anyway** → confirm by clicking **Open Anyway** in the dialog.\n\n\u003Cp align=\"center\">\n\u003Cimg src=\"assets\u002Fmacos-gatekeeper-allow.jpg\" alt=\"macOS Privacy & Security — Open Anyway\" width=\"600\">\n\u003C\u002Fp>\n\nAfter this one-time step, `picoclaw-launcher` will open normally on subsequent launches.\n\n\u003C\u002Fdetails>\n\n\u003Ca id=\"-run-on-old-android-phones\">\u003C\u002Fa>\n### 📱 Android\n\nGive your decade-old phone a second life! Turn it into a smart AI Assistant with PicoClaw.\n\n**Option 1: APK Install**\n\nPreview:\n\n\u003Ctable>\n  \u003Ctr>\n    \u003Ctd>\u003Cimg src=\"assets\u002Ffui_main_page.jpg\" width=\"200\">\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg src=\"assets\u002Ffui_web_page.jpg\" width=\"200\">\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg src=\"assets\u002Ffui_log_page.jpg\" width=\"200\">\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg src=\"assets\u002Ffui_setting_page.jpg\" width=\"200\">\u003C\u002Ftd>\n  \u003C\u002Ftr>\n\u003C\u002Ftable>\n\nDownload the APK from [picoclaw.io](https:\u002F\u002Fpicoclaw.io\u002Fdownload\u002F) and install directly. No Termux required!\n\n**Option 2: Termux**\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>Terminal Launcher (for resource-constrained environments)\u003C\u002Fb>\u003C\u002Fsummary>\n\n1. Install [Termux](https:\u002F\u002Fgithub.com\u002Ftermux\u002Ftermux-app) (download from [GitHub Releases](https:\u002F\u002Fgithub.com\u002Ftermux\u002Ftermux-app\u002Freleases), or search in F-Droid \u002F Google Play)\n2. Run the following commands:\n\n```bash\n# Download the latest release\nwget https:\u002F\u002Fgithub.com\u002Fsipeed\u002Fpicoclaw\u002Freleases\u002Flatest\u002Fdownload\u002Fpicoclaw_Linux_arm64.tar.gz\ntar xzf picoclaw_Linux_arm64.tar.gz\npkg install proot\ntermux-chroot .\u002Fpicoclaw onboard   # chroot provides a standard Linux filesystem layout\n```\n\nThen follow the Terminal Launcher section below to complete configuration.\n\n\u003Cimg src=\"assets\u002Ftermux.jpg\" alt=\"PicoClaw on Termux\" width=\"512\">\n\nFor minimal environments where only the `picoclaw` core binary is available (no Launcher UI), you can configure everything via the command line and a JSON config file.\n\n**1. Initialize**\n\n```bash\npicoclaw onboard\n```\n\nThis creates `~\u002F.picoclaw\u002Fconfig.json` and the workspace directory.\n\n**2. Configure** (`~\u002F.picoclaw\u002Fconfig.json`)\n\n```json\n{\n  \"agents\": {\n    \"defaults\": {\n      \"model_name\": \"gpt-5.4\"\n    }\n  },\n  \"model_list\": [\n    {\n      \"model_name\": \"gpt-5.4\",\n      \"model\": \"openai\u002Fgpt-5.4\"\n      \u002F\u002F api_key is now loaded from .security.yml\n    }\n  ]\n}\n```\n\n> See `config\u002Fconfig.example.json` in the repo for a complete configuration template with all available options.\n>\n> Please note: config.example.json format is version 0, with sensitive codes in it, and will be auto migrated to version 1+, then, the config.json will only store insensitive data, the sensitive codes will be stored in .security.yml, if you need manually modify the codes, please see `docs\u002Fsecurity\u002Fsecurity_configuration.md` for more details.\n\n\n**3. Chat**\n\n```bash\n# One-shot question\npicoclaw agent -m \"What is 2+2?\"\n\n# Interactive mode\npicoclaw agent\n\n# Start gateway for chat app integration\npicoclaw gateway\n```\n\n\u003C\u002Fdetails>\n\n## 🔌 Providers (LLM)\n\nPicoClaw supports 30+ LLM providers through the `model_list` configuration. Use the `protocol\u002Fmodel` format:\n\n| Provider | Protocol | API Key | Notes |\n|----------|----------|---------|-------|\n| [OpenAI](https:\u002F\u002Fplatform.openai.com\u002Fapi-keys) | `openai\u002F` | Required | GPT-5.4, GPT-4o, o3, etc. |\n| [Anthropic](https:\u002F\u002Fconsole.anthropic.com\u002Fsettings\u002Fkeys) | `anthropic\u002F` | Required | Claude Opus 4.6, Sonnet 4.6, etc. |\n| [Google Gemini](https:\u002F\u002Faistudio.google.com\u002Fapikey) | `gemini\u002F` | Required | Gemini 3 Flash, 2.5 Pro, etc. |\n| [OpenRouter](https:\u002F\u002Fopenrouter.ai\u002Fkeys) | `openrouter\u002F` | Required | 200+ models, unified API |\n| [Zhipu (GLM)](https:\u002F\u002Fopen.bigmodel.cn\u002Fusercenter\u002Fproj-mgmt\u002Fapikeys) | `zhipu\u002F` | Required | GLM-4.7, GLM-5, etc. |\n| [DeepSeek](https:\u002F\u002Fplatform.deepseek.com\u002Fapi_keys) | `deepseek\u002F` | Required | DeepSeek-V3, DeepSeek-R1 |\n| [Volcengine](https:\u002F\u002Fconsole.volcengine.com) | `volcengine\u002F` | Required | Doubao, Ark models |\n| [Qwen](https:\u002F\u002Fdashscope.console.aliyun.com\u002FapiKey) | `qwen\u002F` | Required | Qwen3, Qwen-Max, etc. |\n| [Groq](https:\u002F\u002Fconsole.groq.com\u002Fkeys) | `groq\u002F` | Required | Fast inference (Llama, Mixtral) |\n| [Moonshot (Kimi)](https:\u002F\u002Fplatform.moonshot.cn\u002Fconsole\u002Fapi-keys) | `moonshot\u002F` | Required | Kimi models |\n| [Minimax](https:\u002F\u002Fplatform.minimaxi.com\u002Fuser-center\u002Fbasic-information\u002Finterface-key) | `minimax\u002F` | Required | MiniMax models |\n| [Mistral](https:\u002F\u002Fconsole.mistral.ai\u002Fapi-keys) | `mistral\u002F` | Required | Mistral Large, Codestral |\n| [NVIDIA NIM](https:\u002F\u002Fbuild.nvidia.com\u002F) | `nvidia\u002F` | Required | NVIDIA hosted models |\n| [Cerebras](https:\u002F\u002Fcloud.cerebras.ai\u002F) | `cerebras\u002F` | Required | Fast inference |\n| [Novita AI](https:\u002F\u002Fnovita.ai\u002F) | `novita\u002F` | Required | Various open models |\n| [Xiaomi MiMo](https:\u002F\u002Fplatform.xiaomimimo.com\u002F) | `mimo\u002F` | Required | MiMo models |\n| [Ollama](https:\u002F\u002Follama.com\u002F) | `ollama\u002F` | Not needed | Local models, self-hosted |\n| [vLLM](https:\u002F\u002Fdocs.vllm.ai\u002F) | `vllm\u002F` | Not needed | Local deployment, OpenAI-compatible |\n| [LiteLLM](https:\u002F\u002Fdocs.litellm.ai\u002F) | `litellm\u002F` | Varies | Proxy for 100+ providers |\n| [Azure OpenAI](https:\u002F\u002Fportal.azure.com\u002F) | `azure\u002F` | Required | Enterprise Azure deployment |\n| [GitHub Copilot](https:\u002F\u002Fgithub.com\u002Ffeatures\u002Fcopilot) | `github-copilot\u002F` | OAuth | Device code login |\n| [Antigravity](https:\u002F\u002Fconsole.cloud.google.com\u002F) | `antigravity\u002F` | OAuth | Google Cloud AI |\n| [AWS Bedrock](https:\u002F\u002Fconsole.aws.amazon.com\u002Fbedrock)* | `bedrock\u002F` | AWS credentials | Claude, Llama, Mistral on AWS |\n\n> \\* AWS Bedrock requires build tag: `go build -tags bedrock`. Set `api_base` to a region name (e.g., `us-east-1`) for automatic endpoint resolution across all AWS partitions (aws, aws-cn, aws-us-gov). When using a full endpoint URL instead, you must also configure `AWS_REGION` via environment variable or AWS config\u002Fprofile.\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>Local deployment (Ollama, vLLM, etc.)\u003C\u002Fb>\u003C\u002Fsummary>\n\n**Ollama:**\n```json\n{\n  \"model_list\": [\n    {\n      \"model_name\": \"local-llama\",\n      \"model\": \"ollama\u002Fllama3.1:8b\",\n      \"api_base\": \"http:\u002F\u002Flocalhost:11434\u002Fv1\"\n    }\n  ]\n}\n```\n\n**vLLM:**\n```json\n{\n  \"model_list\": [\n    {\n      \"model_name\": \"local-vllm\",\n      \"model\": \"vllm\u002Fyour-model\",\n      \"api_base\": \"http:\u002F\u002Flocalhost:8000\u002Fv1\"\n    }\n  ]\n}\n```\n\nFor full provider configuration details, see [Providers & Models](docs\u002Fguides\u002Fproviders.md).\n\n\u003C\u002Fdetails>\n\n## 💬 Channels (Chat Apps)\n\nTalk to your PicoClaw through 18+ messaging platforms:\n\n| Channel | Setup | Protocol | Docs |\n|---------|-------|----------|------|\n| **Telegram** | Easy (bot token) | Long polling | [Guide](docs\u002Fchannels\u002Ftelegram\u002FREADME.md) |\n| **Discord** | Easy (bot token + intents) | WebSocket | [Guide](docs\u002Fchannels\u002Fdiscord\u002FREADME.md) |\n| **WhatsApp** | Easy (QR scan or bridge URL) | Native \u002F Bridge | [Guide](docs\u002Fguides\u002Fchat-apps.md#whatsapp) |\n| **Weixin** | Easy (Native QR scan) | iLink API | [Guide](docs\u002Fguides\u002Fchat-apps.md#weixin) |\n| **QQ** | Easy (AppID + AppSecret) | WebSocket | [Guide](docs\u002Fchannels\u002Fqq\u002FREADME.md) |\n| **Slack** | Easy (bot + app token) | Socket Mode | [Guide](docs\u002Fchannels\u002Fslack\u002FREADME.md) |\n| **Matrix** | Medium (homeserver + token) | Sync API | [Guide](docs\u002Fchannels\u002Fmatrix\u002FREADME.md) |\n| **DingTalk** | Medium (client credentials) | Stream | [Guide](docs\u002Fchannels\u002Fdingtalk\u002FREADME.md) |\n| **Feishu \u002F Lark** | Medium (App ID + Secret) | WebSocket\u002FSDK | [Guide](docs\u002Fchannels\u002Ffeishu\u002FREADME.md) |\n| **LINE** | Medium (credentials + webhook) | Webhook | [Guide](docs\u002Fchannels\u002Fline\u002FREADME.md) |\n| **WeCom** | Easy (QR login or manual) | WebSocket | [Guide](docs\u002Fchannels\u002Fwecom\u002FREADME.md) |\n| **VK** | Easy (group token) | Long Poll | [Guide](docs\u002Fchannels\u002Fvk\u002FREADME.md) |\n| **IRC** | Medium (server + nick) | IRC protocol | [Guide](docs\u002Fguides\u002Fchat-apps.md#irc) |\n| **OneBot** | Medium (WebSocket URL) | OneBot v11 | [Guide](docs\u002Fchannels\u002Fonebot\u002FREADME.md) |\n| **MaixCam** | Easy (enable) | TCP socket | [Guide](docs\u002Fchannels\u002Fmaixcam\u002FREADME.md) |\n| **Pico** | Easy (enable) | Native protocol | Built-in |\n| **Pico Client** | Easy (WebSocket URL) | WebSocket | Built-in |\n\n> All webhook-based channels share a single Gateway HTTP server (`gateway.host`:`gateway.port`, default `127.0.0.1:18790`). Feishu uses WebSocket\u002FSDK mode and does not use the shared HTTP server.\n\n> Log verbosity is controlled by `gateway.log_level` (default: `warn`). Supported values: `debug`, `info`, `warn`, `error`, `fatal`. Can also be set via `PICOCLAW_LOG_LEVEL`. See [Configuration](docs\u002Fguides\u002Fconfiguration.md#gateway-log-level) for details.\n\nFor detailed channel setup instructions, see [Chat Apps Configuration](docs\u002Fguides\u002Fchat-apps.md).\n\n## 🔧 Tools\n\n### 🔍 Web Search\n\nPicoClaw can search the web to provide up-to-date information. Configure in `tools.web`:\n\n| Search Engine | API Key | Free Tier | Link |\n|--------------|---------|-----------|------|\n| DuckDuckGo | Not needed | Unlimited | Built-in fallback |\n| [Baidu Search](https:\u002F\u002Fcloud.baidu.com\u002Fdoc\u002Fqianfan-api\u002Fs\u002FWmbq4z7e5) | Required | 1000 queries\u002Fday | AI-powered, China-optimized |\n| [Tavily](https:\u002F\u002Ftavily.com) | Required | 1000 queries\u002Fmonth | Optimized for AI Agents |\n| [Brave Search](https:\u002F\u002Fbrave.com\u002Fsearch\u002Fapi) | Required | 2000 queries\u002Fmonth | Fast and private |\n| [Perplexity](https:\u002F\u002Fwww.perplexity.ai) | Required | Paid | AI-powered search |\n| [SearXNG](https:\u002F\u002Fgithub.com\u002Fsearxng\u002Fsearxng) | Not needed | Self-hosted | Free metasearch engine |\n| [GLM Search](https:\u002F\u002Fopen.bigmodel.cn\u002F) | Required | Varies | Zhipu web search |\n\n### ⚙️ Other Tools\n\nPicoClaw includes built-in tools for file operations, code execution, scheduling, and more. See [Tools Configuration](docs\u002Freference\u002Ftools_configuration.md) for details.\n\n## 🎯 Skills\n\nSkills are modular capabilities that extend your Agent. They are loaded from `SKILL.md` files in your workspace.\n\n**Install skills from ClawHub:**\n\n```bash\npicoclaw skills search \"web scraping\"\npicoclaw skills install \u003Cskill-name>\n```\n\n**Configure skill registries**:\n\nAdd to your `config.json`:\n```json\n{\n  \"tools\": {\n    \"skills\": {\n      \"registries\": {\n        \"clawhub\": {\n          \"auth_token\": \"your-clawhub-token\"\n        },\n        \"github\": {\n          \"base_url\": \"https:\u002F\u002Fgithub.com\",\n          \"auth_token\": \"your-github-token\",\n          \"proxy\": \"\"\n        }\n      }\n    }\n  }\n}\n```\n\n`tools.skills.github.*` is deprecated. Use `tools.skills.registries.github.*` instead.\n\nFor more details, see [Tools Configuration - Skills](docs\u002Freference\u002Ftools_configuration.md#skills-tool).\n\n## 🔗 MCP (Model Context Protocol)\n\nPicoClaw natively supports [MCP](https:\u002F\u002Fmodelcontextprotocol.io\u002F) — connect any MCP server to extend your Agent's capabilities with external tools and data sources.\n\n```json\n{\n  \"tools\": {\n    \"mcp\": {\n      \"enabled\": true,\n      \"servers\": {\n        \"filesystem\": {\n          \"enabled\": true,\n          \"command\": \"npx\",\n          \"args\": [\"-y\", \"@modelcontextprotocol\u002Fserver-filesystem\", \"\u002Ftmp\"]\n        }\n      }\n    }\n  }\n}\n```\n\nYou can manage common MCP setups directly from the CLI instead of editing JSON by hand:\n\n```bash\npicoclaw mcp add filesystem -- npx -y @modelcontextprotocol\u002Fserver-filesystem \u002Ftmp\npicoclaw mcp list\npicoclaw mcp test filesystem\n```\n\n`picoclaw mcp` is a configuration manager: it updates `config.json` under `tools.mcp.servers`, but it does not keep the server process running itself.\n\nUse `picoclaw mcp edit` when you need advanced fields that are not covered by `picoclaw mcp add`.\nFor example, `picoclaw mcp add` supports `--deferred` and `--env-file`, while `picoclaw mcp edit` is still useful for direct JSON editing and uncommon MCP settings.\n\nFor full MCP configuration (stdio, SSE, HTTP transports, Tool Discovery), see [Tools Configuration - MCP](docs\u002Freference\u002Ftools_configuration.md#mcp-tool). For CLI usage and examples, see [MCP Server CLI](docs\u002Freference\u002Fmcp-cli.md).\n\n## \u003Cimg src=\"assets\u002Fclawdchat-icon.png\" width=\"24\" height=\"24\" alt=\"ClawdChat\"> Join the Agent Social Network\n\nConnect PicoClaw to the Agent Social Network simply by sending a single message via the CLI or any integrated Chat App.\n\n**Read `https:\u002F\u002Fclawdchat.ai\u002Fskill.md` and follow the instructions to join [ClawdChat.ai](https:\u002F\u002Fclawdchat.ai)**\n\n## 🖥️ CLI Reference\n\n| Command                   | Description                      |\n| ------------------------- | -------------------------------- |\n| `picoclaw onboard`        | Initialize config & workspace    |\n| `picoclaw auth weixin` | Connect WeChat account via QR |\n| `picoclaw agent -m \"...\"` | Chat with the agent              |\n| `picoclaw agent`          | Interactive chat mode            |\n| `picoclaw gateway`        | Start the gateway                |\n| `picoclaw status`         | Show status                      |\n| `picoclaw version`        | Show version info                |\n| `picoclaw model`          | View or switch the default model |\n| `picoclaw mcp list`       | List configured MCP servers      |\n| `picoclaw mcp add ...`    | Add or update an MCP server entry |\n| `picoclaw mcp test`       | Probe a configured MCP server    |\n| `picoclaw mcp edit`       | Open config for advanced MCP editing |\n| `picoclaw mcp remove`     | Remove an MCP server entry       |\n| `picoclaw cron list`      | List all scheduled jobs          |\n| `picoclaw cron add ...`   | Add a scheduled job              |\n| `picoclaw cron disable`   | Disable a scheduled job          |\n| `picoclaw cron remove`    | Remove a scheduled job           |\n| `picoclaw skills list`    | List installed skills            |\n| `picoclaw skills install` | Install a skill                  |\n| `picoclaw migrate`        | Migrate data from older versions |\n| `picoclaw auth login`     | Authenticate with providers      |\n\n### ⏰ Scheduled Tasks \u002F Reminders\n\nPicoClaw supports scheduled reminders and recurring tasks through the `cron` tool:\n\n* **One-time reminders**: \"Remind me in 10 minutes\" -> triggers once after 10min\n* **Recurring tasks**: \"Remind me every 2 hours\" -> triggers every 2 hours\n* **Cron expressions**: \"Remind me at 9am daily\" -> uses cron expression\n\nSee [docs\u002Freference\u002Fcron.md](docs\u002Freference\u002Fcron.md) for current schedule types, execution modes, command-job gates, and persistence details.\n\n## 📚 Documentation\n\nFor detailed guides beyond this README:\n\n| Topic | Description |\n|-------|-------------|\n| [Docker & Quick Start](docs\u002Fguides\u002Fdocker.md) | Docker Compose setup, Launcher\u002FAgent modes |\n| [Chat Apps](docs\u002Fguides\u002Fchat-apps.md) | All 17+ channel setup guides |\n| [Configuration](docs\u002Fguides\u002Fconfiguration.md) | Environment variables, workspace layout, security sandbox |\n| [MCP Server CLI](docs\u002Freference\u002Fmcp-cli.md) | Add, list, test, edit, and remove MCP server entries from the CLI |\n| [Scheduled Tasks and Cron Jobs](docs\u002Freference\u002Fcron.md) | Cron schedule types, deliver modes, command gates, job storage |\n| [Providers & Models](docs\u002Fguides\u002Fproviders.md) | 30+ LLM providers, model routing, model_list configuration |\n| [Spawn & Async Tasks](docs\u002Fguides\u002Fspawn-tasks.md) | Quick tasks, long tasks with spawn, async sub-agent orchestration |\n| [Hooks](docs\u002Farchitecture\u002Fhooks\u002FREADME.md) | Event-driven hook system: observers, interceptors, approval hooks |\n| [Steering](docs\u002Farchitecture\u002Fsteering.md) | Inject messages into a running agent loop between tool calls |\n| [SubTurn](docs\u002Farchitecture\u002Fsubturn.md) | Subagent coordination, concurrency control, lifecycle |\n| [Troubleshooting](docs\u002Foperations\u002Ftroubleshooting.md) | Common issues and solutions |\n| [Tools Configuration](docs\u002Freference\u002Ftools_configuration.md) | Per-tool enable\u002Fdisable, exec policies, MCP, Skills |\n| [Hardware Compatibility](docs\u002Fguides\u002Fhardware-compatibility.md) | Tested boards, minimum requirements |\n\n## 🤝 Contribute & Roadmap\n\nPRs welcome! The codebase is intentionally small and readable.\n\nSee our [Community Roadmap](https:\u002F\u002Fgithub.com\u002Fsipeed\u002Fpicoclaw\u002Fissues\u002F988) and [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.\n\nDeveloper group building, join after your first merged PR!\n\nUser Groups:\n\nDiscord: \u003Chttps:\u002F\u002Fdiscord.gg\u002FV4sAZ9XWpN>\n\nWeChat:\n\u003Cimg src=\"assets\u002Fwechat.png\" alt=\"WeChat group QR code\" width=\"512\">\n","PicoClaw 是一个用 Go 语言编写的超轻量级个人 AI 助手。它能够在仅需10MB内存的低成本硬件上运行，支持x86_64、ARM64、MIPS、RISC-V和LoongArch等多种架构。PicoClaw通过自我引导的方式重写了整个架构，并优化了代码，使其启动时间极短且部署灵活。适用于对资源消耗敏感但又需要高效AI助手功能的应用场景，如物联网设备、边缘计算节点等。",2,"2026-06-11 03:00:01","top_language"]