Your First Project in Claude Code
Zero to shipping in one session
Claude Code is not a chatbot with a code plugin. It is an agent that operates directly in your codebase — reading files, running commands, editing code, and reasoning about architecture. The first time you use it properly, it feels like working with a senior engineer who never gets tired and never takes shortcuts without telling you. This guide gets you there.
What you need
An Anthropic account
Claude Code is available at claude.ai/code or as a CLI. The CLI is more powerful for development work.
Node.js installed
The Claude Code CLI requires Node.js. Run node --version to confirm you have it. If not, install via nodejs.org.
A project directory
Claude Code works in the context of your actual files. Have a real project directory ready, even if it is empty to start.
A clear starting point
You do not need a spec. You need a sentence. "I want to build a Next.js portfolio site with a contact form." That is enough to start.
Your first session
Open Claude Code in your project directory. The most important thing to understand: Claude Code reads your project context automatically. It knows what files exist, what stack you are using, and what conventions are in place.
Start with context-setting, not task-giving. Before you ask it to build anything, tell it what you are trying to do and any constraints that matter. A CLAUDE.md file in your project root is the best way to do this — Claude reads it at the start of every session.
Your first CLAUDE.md should cover: what the project is, the tech stack, code style preferences, and any decisions already made. Claude Code will honor these throughout the session.
Example first message
I want to build a personal portfolio site. Stack is Next.js 15 with TypeScript and Tailwind. I want a hero section, a work section with case study cards, an about section, and a contact form. Start by scaffolding the project structure and the main page layout.
How to work with it
The biggest mistake new users make is treating Claude Code like a code generator. You paste in a request, it produces code, you copy it somewhere. That workflow misses the point entirely.
Claude Code is a collaborator. It maintains context across your entire session. Ask it to build something, review what it built, push back on what you do not like, ask it to explain decisions you do not understand. The back-and-forth is where the value is.
Read files before editing them — Claude does this automatically, but you can ask it to confirm its understanding first
Ask it to explain its architectural decisions before accepting them
If output is not what you expected, describe specifically what is wrong rather than starting over
Use ! prefix to run shell commands directly in the session (e.g., ! npm run dev)
Review diffs before approving edits to files you care about
What to expect
First sessions are often disorienting. Claude Code moves fast and makes decisions without always asking for approval. That is by design — it is trying to be useful, not cautious. You will develop a feel for when to let it run and when to ask it to pause and explain.
Expect to course-correct. The agent will occasionally make assumptions you disagree with. This is normal. Correct it clearly and specifically. It will adjust.
By the end of a focused 2-hour session, most people have shipped something real. That pace is the point. If you are moving slower than that, the bottleneck is usually prompt clarity or scope management, not the tool.
Need help putting this into practice?
Sprintt helps teams implement AI strategy, build product operating models, and ship faster. Reach out for a consultation.