Understand the FULL Claude Ecosystem in One Video

TL;DR: Claude has evolved from a simple chat tool into an AI platform with four main products — Chat, Co‑work, Code, and Design — plus a complete ecosystem of skills, connectors, plugins, projects, memory, and autonomous features. It helps you think, delegate tasks, build applications, and create designs. ## Three Core Products: Chat, Co‑work, Code All three products are included in a single paid subscription. A simple framework helps you quickly tell them apart: - **Chat** → to think - **Co‑work** → to delegate - **Code** → to build ### Chat: A Conversational Thinking Partner The most familiar product. You ask a question or give an instruction, Claude replies, and you iterate back and forth — it’s a conversation. Best for brainstorming, writing, research, and working through ideas. Chat syncs across mobile, web, and desktop, so you can start on your phone and continue on your computer. ### Co‑work: Hand Off the Task Co‑work runs directly on your computer, pointing to a folder as a workspace. Claude starts doing things proactively: reading emails, drafting replies, organizing research materials, building documents and presentations, tidying up files. You give it a task and walk away; when you come back, the work is done. For example, ask it to clean up a messy desktop, and it will analyze all files and images, organizing them into clearly structured folders. Advanced use: you have a meeting tomorrow; Co‑work can find it on your calendar, read the notes, build a presentation, and draft a follow-up email ready to send. Co‑work = Claude + to‑do list + execution ability. ### Code: Build Apps with Natural Language Claude writes all the code for you to create apps, websites, automation tools, Chrome extensions, and more. You describe what you want entirely in natural language, and it builds it. Then you ask for changes and add new features in plain English. A complete beginner can describe a game, internal tool, or dashboard and have a working thing before lunch. Now you are the developer. ### Four in One: Desktop App + Design The desktop app gives you access to all three products at once, each in its own tab. There’s also a fourth product, **Design** (available from the sidebar or at claude.ai/design). Give it a prompt, a codebase, a Figma file, or brand assets, and it returns working prototypes, slides, landing pages, even animations. Once the design is ready, you can hand it directly to Code for deployment. In short: Chat helps you think, Co‑work lets you delegate, Code lets you build, and Design lets you create. ## Feature Layers That Extend the Core Products The following features span Chat, Co‑work, and Code, and some can be found in the “Customize” menu. ### Skills: Reusable Expertise Packages Skills are a set of instructions and guidelines that help Claude perform tasks in a repeatable, consistent way. Many skills are built in — for example, creating a PowerPoint automatically triggers a corresponding skill so you get a well‑designed, editable presentation on the first try. You can also create your own skills: after going back and forth with Claude, say “Package what we just did into a skill,” and it will build one. From then on, every time you do the same kind of task, Claude automatically invokes that skill and carries out all the steps the same way. For instance, a skill that writes in your voice and creates visuals according to your brand style — you don’t have to repeat background and custom prompts each time. This skill can draft content in Chat, automate content creation in Co‑work, or be built into a custom tool in Code. Essentially, a skill is a folder of instructions and guidelines that Claude reads first. ### Connectors: Link Your Everyday Tools Connectors link Claude to the tools you use: Gmail, Slack, Google Drive, Asana, and hundreds of others. Search for a tool name, click to connect and log in, and Claude can act through or pull data from that tool. You control what it can do via a permissions list. Connectors work across all three products too. ### Plugins: Skills + Connectors Bundled Together Plugins bundle skills and connectors around a specific role or tool, plus extra instructions. Anthropic has a full plugin marketplace. For example, the Figma plugin includes skills for correct formatting and a connector to pass information. A legal plugin comes pre‑loaded with skills for drafting briefs, contracts, organizing research, and relevant connectors. If you feel a part of your workflow is repetitive, check if there’s a plugin for it first. ## Context & Memory Layer ### Projects Projects are folders that organize related chats. Inside each project, you can add custom instructions to guide Claude’s behavior, upload files for extra context, and there’s built‑in memory (refreshed automatically). Claude remembers all conversations in that project and learns more over time. For example, a full marketing project with custom instructions for writing style and tone, brand guidelines, an uploaded customer list, past campaign results, and a content calendar. Every time you open the project, Claude already knows all this — you don’t have to start from scratch. Note: projects in Chat and projects in Co‑work are separate, but you can import. ### Memory - **Project‑level memory**: Claude updates it daily automatically and carries it across all chats in that project. - **Account‑level memory**: Applies only to Chat and spans your entire account. Over time, Claude learns your preferences, working style, and life/work context. - **CLAUDE.md file in Claude Code**: In the project folder, side by side with code. The `/init` command makes Claude analyze the folder and conversation content, create the file, save persistent context, and update automatically. Some people also manually create a CLAUDE.md for Co‑work projects. So: projects give persistent context, memory learns who you are, and CLAUDE.md briefs Code/Co‑work on how to work. ## Autonomy & Automation Layer ### Computer Use Claude can control your desktop: open apps, browse the browser, fill in forms, move files. You just give it a task, and it operates the computer to do it. Still early, but it shows the future direction. ### Dispatch Lets you control a Co‑work session from your phone. You can start tasks, check status, or redirect from anywhere. ### Scheduled Tasks Currently a notable feature in Co‑work. Set it once, and Claude runs automatically on a schedule. For example, a morning briefing: check the calendar, scan your inbox, pull items from connected tools, and have a full summary ready before your first meeting. Also useful for weekly competitor research reports, daily social media drafts, daily summaries, etc. Anything you manually trigger regularly can be handed to scheduled tasks. ### Claude Code Routines The Code version. You can save a configuration of prompts, repositories, and connectors, set it to run automatically without your computer staying on. Triggers can be schedules, GitHub events, or API calls. Best for experienced Claude Code users. ## Advanced Features & Summary Here’s the complete picture of the Claude ecosystem: - **Four Products**: Chat (think), Co‑work (delegate), Code (build), Design (create) - **Common Layer**: Skills (repetition), Connectors (link tools), Plugins (bundles) - **Context & Memory Layer**: Projects (persistent work), Memory (learns you), CLAUDE.md (briefing) - **Autonomy Layer**: Computer Use (control desktop), Dispatch (remote Co‑work), Scheduled Tasks (automate Co‑work), Routines (auto‑trigger code runs) Each product has deeper videos available — check the Claude ecosystem playlist. Source: https://www.youtube.com/watch?v=MS7E5TXNviM