Understand the FULL Claude Ecosystem in One Video

Overview

Claude’s “ecosystem” has expanded from a simple chat box into multiple products plus an automation/context layer—enough that it can feel overwhelming. The goal is to explain:

• What each product does
• How they differ
• When to use them
• The features that expand their capabilities

---

Core Products (3 flagship tools + Design)

Chat (“think”)

Best for interactive work:

• Discussion loops, writing, research, and working through ideas
• Works like an ongoing conversation where you iterate with Claude
• Syncs across mobile, web, and desktop so you can continue where you left off
• Uses the same paid subscription access as the other flagship products

Co-work (“delegate”)

Focused on delegating tasks rather than walking through every step:

• Runs on desktop and uses a folder/workspace you point it to
• Claude performs actions such as:
  • Drafting emails/replies
  • Pulling together research
  • Organizing files
  • Building documents and presentations
• Example uses:
  • Organizing a messy desktop
  • Pulling meeting info + notes, then creating a presentation and follow-up email

Code (“build”)

Focused on building software/automation using natural language:

• Claude writes the actual code for:
  • Apps
  • Websites
  • Automations
  • Chrome extensions
  • Productivity tools
• Emphasis: you don’t need to be a developer—just describe changes/features in plain English

Design (“create”) (newer, more separate but recommended)

Used for prototyping and producing design outputs:

• You can provide prompts, a codebase, a Figma file, and brand assets
• Claude returns working design deliverables such as:
  • A prototype
  • Slide deck
  • One-pager
  • Landing page
  • Animation
• After designing, you can hand it off directly to Code to ship

---

Simple Mental Model Across Products

• Chat = think (conversation/iteration)
• Co-work = delegate (automate task completion in the background)
• Code = build (generate implementations)
• Design = create (generate design/prototypes)

---

Desktop Access Model

• The desktop app is positioned as the easiest way to use all three together:
  • Chat / Co-work / Code appear in adjacent tabs
• Design is accessed via:
  • The Claude sidebar, or
  • claude.ai/design

---

Free Guide Mentioned (Tutorial / Resources)

A promoted free guide (linked in the description) includes four sections:

1. Getting started
2. Core features
3. Real-world applications
4. Advanced implementation

It includes use cases like:

• Content creation
• Data analysis
• Strategic planning

…and provides frameworks and ready-to-use prompts.

---

“Universal Layer” Features Across Chat / Co-work / Code

These plug into the core products to repeat work, connect tools, and package workflows.

Skills

• Expertise packages for repeatable, consistent execution
• Built-in example:
  • Claude reads a PowerPoint skill first to format correctly
• You can create your own skill by packaging prior successful iterations (e.g., “package what we went through as a skill”)
• Skills can carry preferences and instructions (e.g., writing in your voice, brand visuals)
• Reusable in Chat, Co-work, and Code

Connectors

• Link Claude to tools you already use (examples: Gmail, Slack, Google Drive, Asana, and more)
• Typical workflow:
  • Search connector → connect → sign in → Claude can take actions/pull data
• Controlled via permissions
• Used across:
  • Chat (Q&A with pulled data)
  • Co-work (automated tool actions)
  • Code (custom personalized tools)

Plugins

• Bundle skills + connectors around a specific role/tool, with extra instructions
• Example plugin types:
  • Figma plugin: design workflow bridging (format-aware creation + passing information)
  • Legal plugin: drafting briefs/contracts, organizing research, managing knowledge, plus connectors
• Suggested approach:
  • If work is repetitive, look for a plugin rather than rebuilding

---

Context & Memory Layer (Improves Usefulness Over Time)

Projects

• Organize related chats (and co-work)
• Add custom instructions and uploaded files as context
• Includes built-in memory that refreshes automatically, so Claude remembers across conversations within that project

Example: a marketing project might include:

• Brand voice
• Guidelines
• Skills to use
• Clients
• Past campaign results
• Content calendar

Important limitation:

• Chat projects and Co-work projects are separate and don’t sync
• You can import a chat project into co-work

Two types of memory described:

• Project memory: automatic within the project
• Account-level memory: preferences/work style, but only applies to Chat (not Co-work/Code)

Memory in Code via cloud.md

• Code uses a persistent cloud.md file as the context “brief”
• Proposed workflow:
  • After creating a project, run /init
  • This analyzes the folder and conversations to create/update cloud.md
• cloud.md becomes a project brief that helps Claude build consistently

---

Autonomous / Automation Layer (Claude Acts Without Prompting)

Computer use

Claude can take control of the desktop:

• Open apps and navigate browsers
• Fill forms
• Move files

It’s framed as early/experimental (“you will run into some issues”).

Dispatch

• Control a Co-work session from a phone:
  • Start a task
  • Check in
  • Redirect

Scheduled tasks (highlighted as most important right now for Co-work)

• Set tasks once and have them run automatically on a schedule
• Examples:
  • Morning brief: check calendar + scan email + pull connected-tool info → deliver summary before meetings
  • Weekly competitive research report
  • Daily social media drafts
  • End-of-day team summaries

Cloud Code routines

• Automate code-related workflows even when your computer is off
• Routines save configurations of prompts, repositories, and connectors, then execute via:
  • schedules
  • GitHub events
  • API calls
• Positioned as more for experienced Code users

---

Recap (Full Ecosystem Structure)

• 4 products:
  • Chat (think), Co-work (delegate), Code (build), Design (create)
• Universal layer:
  • Skills (repeat), Connectors (link tools), Plugins (bundle roles)
• Context/memory layer:
  • Projects (persist + refresh memory), cloud.md for Code
• Autonomous layer:
  • Computer use (desktop control), Dispatch (remote Co-work control), Scheduled tasks (automatic Co-work), Routines (trigger code automatically)

---

Main Speakers / Sources

• Speaker: the video’s narrator/creator (referenced via phrases like “I’m breaking down…”, “I have videos going deeper…”, “free guide… click the link…”)
• Referenced organization: Anthropic (mentioned as hosting a plugin marketplace)