OpenClaw
An AI agent tool built around a skill catalog called ClawHub. Released in November 2025, it features over 137 version iterations and more than 5,400 skills, with support for tool invocation, memory management, and gateway functionality.
Core Facts
Limited by the computing power and memory of consumer-grade devices, the scale of models that can be run locally is typically far smaller than cloud-based flagship models, with gaps in complex reasoning, long-context processing, and other tasks.
A Skill is a modular encapsulation that grants an agent specialized capabilities, containing complete workflows, judgment logic, and contextual knowledge, and can combine multiple tools to accomplish complex goals.
OpenClaw is an open-source AI Agent framework with tool-calling, memory management, and autonomous evolution capabilities.
Timeline (last 90 days)
Claude Code and OpenClaw both fall under the category of AI Agents, with the difference lying in differing degrees of autonomy and persistence.
Kimi Claw is an AI product based on OpenClaw that supports creating automated tasks on a computer and features a long- and short-term memory mechanism (automatically recording information via Memory.md every day).
Domestic products of the same kind based on OpenClaw have roughly similar basic capabilities, with the differences mainly in the choice of large model.
WorkBuddy is similar to automation tools such as OpenClaw in how it works, both adopting a "brain + operator" architecture, but WorkBuddy is an out-of-the-box commercial product while OpenClaw is an open-source framework that must be configured by the user.
The UP host judges that domestic and overseas AI Agent products will sooner or later converge, and that the functional boundaries between general-purpose Agents (WorkBuddy, OpenClaw) and development-oriented Agents (Codex, Claude Code, Trae) will gradually blur.
Scenario-specific agents such as Viktor (for finance/operations) and OpenClaw (an open-source agent framework) perform better in niche domains but lack generality.
The AI agent market currently already has multiple products such as Viktor, Lindy, Manus, OpenClaw, Microsoft Copilot Cowork, and Zapier AI agents, and the field is crowded.
OpenClaw's advantage is that it is auditable and modifiable, making it suitable for technically capable teams building vertical-scenario agents; its disadvantage is the lack of an out-of-the-box experience, making it hard for ordinary users to use directly.
WorkBuddy does similar things to agent products such as OpenClaw, but backed by the Tencent ecosystem, it has a natural advantage in connecting with scenarios such as WeChat, WeCom and Tencent Docs.
The open-source version of OpenClaw has two major flaws: it is unsafe (it may mistakenly delete system files or local files) and token consumption is uncontrollable (it does not support subscription plans, and completing a task may cost hundreds or even thousands of yuan).
40 more timeline events
All Facts (20)
Limited by the computing power and memory of consumer-grade devices, the scale of models that can be run locally is typically far smaller than cloud-based flagship models, with gaps in complex reasoning, long-context processing, and other tasks.
90%VerifiedA Skill is a modular encapsulation that grants an agent specialized capabilities, containing complete workflows, judgment logic, and contextual knowledge, and can combine multiple tools to accomplish complex goals.
90%VerifiedOpenClaw is an open-source AI Agent framework with tool-calling, memory management, and autonomous evolution capabilities.
80%VerifiedOpenAI's agents can browse the web, write and execute code, manage files, and complete complex workflows through multi-turn reasoning
75%VerifiedTechnology, tools, and models in the AI field iterate extremely fast, and a guide that has not been updated for half a year may already contain a large amount of outdated content.
65%VerifiedOpenClaw carries security risks: because its built-in tools have extremely high permissions (capable of operating system files, reading browser information, and executing tasks online), any configured API Keys are highly susceptible to leakage if the machine is compromised.
65%VerifiedOpenClaw's architecture is: a person connects to OpenClaw (local computer) via communication software (WhatsApp, Telegram, etc.), which then connects to a language model (cloud/local); OpenClaw itself is an interface program between the person and the language model and does not possess independent intelligence.
65%Unverifiedopenclaw gateway status 命令用于检查OpenClaw Gateway的运行状态，RPC状态显示OK表示服务正常运行
95%UnverifiedOpenClaw使用Compaction机制解决Context Window不够用的问题，将较旧的对话历史发给语言模型做摘要，用摘要替换原文，该过程可递归进行
90%UnverifiedOpenClaw的心跳机制每隔固定时间（如30分钟）自动触发语言模型，让它读取habit.md中的待办事项并执行，配合CronJob排程系统实现定时任务
90%UnverifiedOpenClaw的记忆检索本质上是RAG（检索增强生成），将记忆文件切成Chunk，通过字面匹配和语义Embedding的加权相似度排序，取Top-K结果返回给语言模型
90%UnverifiedOpenClaw中最强大也最危险的工具是Execute，可以执行任何Shell命令，存在语言模型执行破坏性命令（如rm -rf）的风险
90%Unverifiednotebooklm-py可作为Claude Code、Codex、OpenClaw等AI智能体的工具插件
90%UnverifiedClaude Code和OpenClaw都可以将飞书CLI作为外部工具调用，从而获得操作飞书的能力
90%UnverifiedARIS采用纯Markdown技能系统，不依赖任何特定框架，能与Claude Code、Codex、OpenClaw或任意LLM代理配合
90%UnverifiedMoltbot这个名字只存活了三天（2026-01-27至2026-01-30）就被替换为最终的OpenClaw
90%UnverifiedOpenClaw项目经历了六次命名：Warelay（2025-11-24）→ CLAWDIS（2025-12-03）→ CLAWDBOT（2026-01-04）→ Clawdbot（2026-01-10）→ Moltbot（2026-01-27）→ OpenClaw（2026-01-30）
90%UnverifiedOpenClaw项目最初名为Warelay，于2025年11月24日首次提交，定位为基于Twilio的WhatsApp中继CLI工具
90%Unverifiedsoul.md和user.md配置文件写得过于冗长会在每轮对话开始时预消耗大量token，导致模型能处理的实际对话内容减少
90%UnverifiedOpenClaw的Dashboard只是前端展示层，真正的业务逻辑全部由Gateway驱动
90%