Orkas is an open-source (MIT), local-first AI team for your desktop: a Commander turns your goal into to-dos and coordinates 16 built-in specialist agents plus Claude Code, Codex, OpenCode, OpenClaw and Hermes, with projects, scheduled automations and connectors. Free app for macOS & Windows; built-in models or your own keys.
JavaScript agent-collaborationagent-harnessagent-orchestrationagent-teamagent-workflowai-assistantai-automationai-employeeai-productivityai-teamai-workflowautonomous-agentsclaude-code-guiclaude-coworkdesktop-agentlocal-firstmulti-agentmulti-agent-systemself-evolving-agentstask-automation
docs: update README key points images
wyt3156c7271bbf
diff --git a/README.md b/README.mdindex 467a729b..72b477f3 100644--- a/README.md+++ b/README.md@@ -18,40 +18,9 @@----## Four core points+## Core features-### 👥 Multi-agent collaboration--**A whole team in one chat window** — the commander dispatches, agents handle their specialties, you just talk.--- **Smart dispatch** — the commander has the full conversation context and decides on its own who to bring in and when, based on your needs and each agent's strengths-- **Orchestrated collaboration** — multiple agents work in serial or in parallel within the same chat; task breakdown, hand-offs between agents, and result aggregation are all orchestrated by the commander-- **Quickly create new agents** — describe one to the commander, or have it distill one from your past chats — it produces a reusable agent you can summon next time-- **Step in anytime** — `@` any member to add requirements, redirect, or pull someone in / kick them out--### 🌱 Self-evolving agents--**Agents that get to know your work better the more you use them** — after each task, an agent reflects on how it did and how to do better next time, and accumulates experience over time.--- **Reflective evolution** — after self-reflection, an agent updates its own playbook: what it's good at / not good at, and which methods work in which situations-- **Skill crystallization** — the moves that worked this time get crystallized into a reusable skill, so the next similar task picks it up directly-- **Private knowledge** — each agent has its own private skill library and memory, never crossed over from other agents--### 💾 Local storage--**Conversations, files, API keys, knowledge bases, custom agents / skills, memory — all stay on your machine.**--- **Offline-capable** — except for the moment of calling the model API, everything else still runs with the network off-- **Config is just files** — all data is plain files: readable, backupable, syncable through your own cloud drive; migrating machines is just copying a directory--### 🖥️ Desktop app--**As a desktop app, local file handling is smooth and the GUI beats the command line.**--- **Direct local file access** — drop files into the chat as attachments; agents can read and write files in your workspace, run scripts, and generate PDFs / images / code; produced files appear as cards in the chat, one click to reveal in Finder / Explorer-- **Visual operations** — agents, skills, and knowledge bases are all managed visually; view images, videos, and generated documents directly in the chat without switching to a CLI-- **Use your own subscription or API key** — log in via OAuth or paste an API key for DeepSeek · Kimi · GLM · MiniMax · Doubao · Qwen · Claude · OpenAI · Gemini; requests don't go through Orkas and aren't archived-- **Cross-platform** — macOS (Apple Silicon + Intel), Windows, and Linux+---diff --git a/README.zh-CN.md b/README.zh-CN.mdindex 47672cbf..20afe613 100644--- a/README.zh-CN.md+++ b/README.zh-CN.md@@ -18,40 +18,9 @@----## 四个核心点+## 核心功能-### 👥 多智能体协作--**一个对话框里坐着一支团队**——指挥官负责调度,智能体各做专长,你说话就行。--- **智能调度**——指挥官掌握整个对话的上下文,根据你的需求和团队成员的能力,自行决定该召集谁、什么时候召集-- **编排协作**——多个智能体在同一对话里串联或并行工作;任务拆解、智能体之间的衔接、结果汇总,全由指挥官来编排-- **快速创建智能体**——根据你的描述,或者直接从过往对话记录里提炼,指挥官会快速生成一个可复用的智能体,下次直接召唤-- **随时介入**——你可以 `@` 任何成员追加要求、重定向,或者把谁拉进来 / 踢出去--### 🌱 智能体自我演进--**让智能体越用越懂你的活**——每次任务跑完,智能体会自动复盘做得怎样、下次怎么做更好,把经验沉淀下来。--- **复盘进化**——智能体自我反思后会更新自己的工作方式:擅长 / 不擅长什么、什么场景下用什么方法-- **沉淀技能**——这次解决问题用对了的招式,智能体会把它固化成一个可复用的技能,下次同类任务直接拿来用-- **专属知识**——每个智能体有自己的私有技能库与记忆,不会被其它智能体串扰--### 💾 本地存储--**对话、文件、API key、知识库、自定义智能体 / 技能、记忆——全部留在你的机器上。**--- **离线可用**——除了调模型 API 那一刻必须联网,其它时候关掉网都能跑-- **配置即文件**——所有数据都是普通文件,可读、可备份、可接你自己的网盘做同步,迁移机器只是拷贝目录--### 🖥️ 桌面应用--**作为桌面应用,本地文件处理顺手,图形界面比命令行友好。**--- **本地文件直通**——拖拽文件进对话即附件;智能体能直接在你的工作区里读写文件、跑脚本、生成 PDF / 图片 / 代码;产出的文件在对话里以卡片呈现,一键在 Finder / 资源管理器中打开-- **可视化操作**——智能体、技能、知识库全部图形化管理;对话里直接看图片、视频、生成的文档,不用切命令行-- **用自己的会员或 API key 接模型**——OAuth 登录会员账号或填 API key,接入 DeepSeek · Kimi · GLM · MiniMax · 豆包 · Qwen · Claude · OpenAI · Gemini 等;请求 Orkas 不经手、不存档-- **跨平台**——macOS(Apple Silicon + Intel)、 Windows、Linux+---diff --git a/resources/key-points-en.png b/resources/key-points-en.pngnew file mode 100644index 00000000..5e82ef94Binary files /dev/null and b/resources/key-points-en.png differdiff --git a/resources/key-points-zh.png b/resources/key-points-zh.pngnew file mode 100644index 00000000..ca57b0f8Binary files /dev/null and b/resources/key-points-zh.png differ