Setup Claude Code
Setup Claude Code
This guide walks you through installing Claude Code, setting up the project we’ll build throughout this series, and running it locally. Follow every step carefully — by the end, your environment will be fully ready.
This is the foundation episode of the series. Everything we build later depends on the setup done here. Don’t skip any step.
🧩 What is Claude Code?
Claude Code is a terminal-based AI coding agent built by Anthropic. Unlike Copilot or Cursor, Claude Code doesn’t just complete lines — it reads your entire project, understands the codebase, and executes multi-step tasks autonomously using tools like file editing, Bash execution, and web search.
Claude Code runs directly in your terminal and integrates with any editor — VS Code, Neovim, or even just the terminal itself.
💳 Pricing — What You Need to Know
Claude Code is paid software. It uses Anthropic’s proprietary models (Sonnet, Opus, Haiku) under the hood.
| Plan | Cost (approx.) | Best For |
|---|---|---|
| Claude Pro | ₹2,000 / month | Full Claude Code access |
| Ollama Cloud Models | Free tier available | Students & hobbyists |
| Ollama Local Models | Free (your hardware) | Offline, privacy-first use |
Students: Don’t want to spend ₹2,000? There’s a free workaround using Ollama — covered at the end of this guidec or they can use
openrouterfree model with claudw code.
🚀 Step 1 — Create a Claude Account & Subscribe
Before installing Claude Code, you need an active Anthropic account with an upgraded plan.
- Go to claude.ai
- Create an account (or log in)
- Navigate to Upgrade Plan
- Choose and confirm the subscription
Once subscribed, you’re authorized to use Claude Code.
📦 Step 2 — Install Claude Code
Installation is a single command. Open your Terminal (macOS/Linux) or Command Prompt (Windows):
On my machine i’ll use below cammand which is for mac os.
curl -fsSL https://claude.ai/install.sh | bash
After installation, verify it worked:
claude --version
Claude Code requires Node.js 18+. If you get an error, install or update Node.js first.
▶️ Step 3 — Run Claude Code
Navigate to any directory and run:
claude
The first time, Claude will ask:
Do you trust the files in this folder?
Since it’s your own machine, select Yes.
If prompted to log in, you’ll be redirected to your browser to Authorize your Anthropic account. Click Authorize and return to the terminal.
Authorization only happens once. Subsequent runs start immediately.
🗂️ Step 4 — Set Up the Project
We’ll work on a pre-built Flask project called Spendly — a personal expense tracker. This simulates a real-world scenario where you join a project with an existing codebase.
Download the Project
Download the project ZIP from here. Extract it and open the folder in VS Code.
Open the Terminal Inside VS Code
Open the integrated terminal in VS Code and navigate to the project folder. Then launch Claude from inside the project:
Setup Commands
Run these in order inside Bash mode:
# 1. Create a virtual environment
uv .venv
# 2. Activate it
source .venv/bin/activate # macOS / Linux
# venv\Scripts\activate # Windows
# 3. Install dependencies
uv pip install -r requirements.txt
# 4. Run the Flask app
uv run app.py
To run these above commands in Claude Code, open Claude terminal in the project folder and run these commands in Bash mode. For Bash mode use Shift + 1.
Why Bash mode matters: Commands run inside Claude’s Bash mode are recorded in the session history. That means you can later ask “What libraries were installed?” and Claude already knows — because it saw the commands.
Open your browser and visit:
http://localhost:5001
You should see the Spendly home page with a landing page, registration UI, and login UI. The backend logic isn’t implemented yet — that’s what we’ll build together throughout the playlist.
🤖 Step 6 — Explore Claude’s Project Understanding
One of Claude Code’s most powerful features is its ability to understand an entire codebase instantly. Ask Claude these three questions whenever you start a new project:
Question 1 — What does this project do?
What does this project do?
Claude reads every file, builds context, and gives you an accurate summary.
Question 2 — What is the tech stack?
What tech stack does this project use?
Claude will break it down by layer:
| Layer | Technologies |
|---|---|
| Backend | Python 3.11, Flask, SQLite |
| Frontend | Jinja2 templates, Vanilla CSS, Google Fonts |
| Testing | pytest, pytest-flask |
Question 3 — Explain the project structure
Explain the project structure to me.
Claude returns a detailed file hierarchy — folders, their purpose, key files, and how they connect.
Make these three questions a habit at the start of every new project you work on with Claude Code.
🐙 Step 7 — Push to GitHub
Every feature we build will be committed and pushed to GitHub. This also prepares us for deployment later.
Create a Repository
- Go to github.com
- Click New Repository
- Name it
spendly - Add a description: An expense tracking application built using Claude Code
- Click Create Repository
Initialize Git
Run these commands inside Bash mode in Claude:
# Initialize git
git init
# Stage all files
git add .
# Commit
git commit -m "Initial commit"
# Add the remote (replace with your repo URL)
git remote add origin https://github.com/YOUR_USERNAME/spendly.git
# Push
git push origin main
Refresh GitHub — your project files should now be visible. ✅
🆓 Step 8 — Free Alternative: Using Ollama (For Students)
If you don’t want to pay for Claude Pro, you can use Ollama to run open-source LLMs with Claude Code — either via cloud models (free quota) or entirely locally.
Install Ollama
Visit ollama.com and follow the installation instructions for your OS. Then:
ollama launch claude
Option A — Ollama Cloud Models (Free Quota)
When prompted, choose a cloud model such as:
- Qwen 3.5 (recommended for coding)
- GLM 5
- K2.5
Claude Code will now use that model instead of Anthropic’s servers. You’ll get a free usage quota each month.
Cloud quotas are limited. Once you hit your limit, you’ll need to wait for the monthly reset or switch to a local model.
Option B — Local Models (Fully Free, No Quota)
Download a local coding model:
# Pull a model locally (choose based on your RAM)
ollama pull qwen2.5-coder:14b # 16 GB RAM or more
ollama pull qwen2.5-coder:7b # 8 GB RAM
| RAM Available | Recommended Model |
|---|---|
| 16 GB+ | qwen2.5-coder:14b |
| 8 GB | qwen2.5-coder:7b |
| Less than 8 GB | qwen2.5-coder:3b |
After download, exit the current session and relaunch:
/exit
ollama launch claude
# Select the local model from the list
Local models are slower than cloud models, especially without a dedicated GPU. Performance improves significantly with more RAM and a CUDA/Metal GPU.
Other Free Alternatives
| Tool | Description |
|---|---|
| OpenRouter | Route Claude Code to multiple AI APIs with free tiers |
| OpenCode | Open-source Claude Code alternative |
Search YouTube for setup guides for each of these.
✅ Summary — Your Setup is Complete
Here’s what you’ve accomplished:
- Installed Claude Code
- Authorized your Anthropic account
- Set up the Spendly project and ran it locally
- Learned how Bash mode works inside Claude
- Pushed the project to GitHub
- Explored Claude’s project understanding capabilities
- Learned about free Ollama alternatives
Quick Start Reference
# Paid version (Anthropic)
claude
# Free version (Ollama)
ollama launch claude
Recommendation: If you can afford ₹2,000/month and want to follow this playlist to its full potential, go with the official Claude subscription. The experience is significantly better. If you’re a student, start with Ollama Cloud — it’s genuinely usable.