.png){kind=small}
The Surfaces of Claude: Where It Lives
A map of every place Claude appears, and which one to reach for when.
What You Will Learn
- Name all the Claude surfaces without looking them up.
- State which surface is best suited for each of five common tasks.
- Explain the difference between interactive and programmatic surfaces.
- Identify which surfaces require which plan tiers.
- Choose the right entry point for a task instead of defaulting to one.
Why This Matters
Most beginners discover one surface — usually claude.ai chat — and use it for everything. That works for a week, then they hit a wall: "why can't Claude see my codebase?", "can I use this from the terminal?", "how do I automate this?". Knowing the surfaces is like knowing the tools in a chef's knife roll.
Concept Explained
A "surface" is a place where Claude appears and where you interact with it. Each surface has its own context model, latency profile, and affordances. Some are interactive (you type, it responds); one is programmatic (you call it via code).
How It Works
All surfaces share the same underlying model — they differ in context assembly and interaction shape: (1) claude.ai chat — conversational, web-based; (2) Projects — knowledge containers with persistent context; (3) Artifacts — interactive shareable content (code, SVGs, documents); (4) Anthropic API — programmatic, used in code; (5) Claude Code — Anthropic's CLI for terminal-based development; (6) MCP servers — expose tools to Claude.
Step-by-Step Tutorial
1. Map each surface to its primary use
For each surface, decide what it is best at. claude.ai chat = quick questions and drafting; Projects = recurring context; Artifacts = interactive content; API = automation; Claude Code = terminal development; MCP = tool integration.
2. Decide interactive vs programmatic
Surfaces 1–3 are interactive (you type, it responds). Surface 4 (API) is programmatic (you call it in code). Surface 5 (Claude Code) is interactive but terminal-based. Knowing this changes how you phrase your request.
3. Check plan availability
Free/Pro: claude.ai chat, basic Projects, basic Artifacts. Pro+: all surfaces. API: pay-per-token, no plan needed. Claude Code: Pro+ or pay-per-token. MCP: works on Pro+.
4. Pick the right surface for five common tasks
Test yourself: (1) write a function → claude.ai chat; (2) write a script that calls Claude → API; (3) debug from terminal → Claude Code; (4) recurring task with shared context → Projects; (5) build a shareable interactive demo → Artifacts.
Real-World Example
A platform team adopted Claude across multiple surfaces. Their first month was chaotic: every team member used claude.ai chat for everything, including multi-file refactors ("why does Claude keep forgetting the types?"). After a one-hour workshop on surfaces, they adopted a simple rule: if a task touches one file, use chat; if it touches multiple files or needs terminal access, use Claude Code; if it needs persistent knowledge, use Projects; if it needs to be automated, use the API. Within two weeks, Claude-induced confusion dropped 60%.
Example Prompts / Commands / Code
Task Surface Why
--------------------------------------------------------------
Draft an email claude.ai chat Quick, conversational
Build a React component demo Artifacts Interactive, shareable
Recurring research on a competitor Projects Persistent knowledge
Automate customer support triage API Programmatic, scalable
Refactor 5-file Python module Claude Code Terminal, multi-file
Connect Claude to internal API MCP Tool integration
Common Mistakes
- Using claude.ai chat for multi-file work — it has no file system access.
- Forgetting Projects exist — re-attaching context every conversation.
- Not knowing Claude Code exists — switching to terminal anyway.
- Trying to use the API for one-off questions — over-engineered.
Best Practices
- Match the surface to the task's scope.
- Default to claude.ai chat; reach for others deliberately.
- Use Projects for recurring context (team knowledge, codebase).
- Use Claude Code for terminal-native development.
- Use the API when you need automation, scale, or integration.
Troubleshooting
| Problem | How to Fix |
|---|---|
| Chat can't see my codebase | Use Claude Code, or attach files explicitly in chat. |
| Context resets every conversation | Use Projects to persist knowledge. |
| Want to automate a Claude task | Use the Anthropic API. |
| Need Claude to call my internal service | Build an MCP server (see PR-23). |
Practical Exercise
Your Turn
For each of these tasks, write down which surface you'd reach for: (1) draft a tweet; (2) build a calculator web app; (3) recurring analysis of weekly metrics; (4) refactor 3 Python files; (5) connect Claude to your team's internal API. Compare with the answer in the examples.
Key Takeaways
- Claude lives on multiple surfaces, not just chat.
- claude.ai chat for quick tasks; Projects for persistent context.
- Artifacts for interactive shareable content.
- API for automation; Claude Code for terminal; MCP for tools.
- Surface selection is a skill — practice deliberately.
Frequently Asked Questions
Which surface should I learn first?
Are all surfaces available on every plan?
Can I use Claude Code without an editor?
What's the difference between Projects and Artifacts?
Further Reading
Official References
Related lessons: BE-01, BE-03, BE-10
SEO Metadata
SEO title: The Surfaces of Claude: Where It Lives
Meta description: A map of every place Claude appears, and which one to reach for when.
Primary keyword: claude surfaces
Secondary keywords: claude.ai, anthropic api, claude code, claude projects, claude artifacts
Search intent: Informational
URL slug: /claude-surfaces-where-it-lives-map
Categories: AI Tools, Claude
Tags: Claude, Beginner, Surfaces, claude.ai, API, Claude Code, IMCSEIAN, Tutorial, IMCSEIAN
Featured image concept: Diagram showing all Claude surfaces arranged around a central developer icon.
Comments
Comments
Post a Comment