← Voltar aos guias

Best Terminal and iTerm2 Setup for Claude Code

1) Start with a clean terminal baseline

Claude Code feels best when your terminal is fast, readable, and predictable. If you are on macOS, iTerm2 is the easiest upgrade over the default terminal because it gives you tabs, split panes, strong search, and better font rendering without getting in your way. Keep the setup simple: one profile, one theme, one font, and no noisy plugins that slow down shell startup.

Use a font made for coding, such as JetBrains Mono, Fira Code, or Meslo. Turn on ligatures only if you like them. Pick a dark theme with strong contrast, because Claude Code output often includes code blocks, diffs, and long commands that are easier to scan when the screen is calm.

• Set the font size to something you can read for long sessions.
• Enable a dark background and clear cursor color.
• Turn off unnecessary animations and visual clutter.

2) Make your shell startup fast

Claude Code is much nicer when your shell opens instantly. Whether you use zsh or bash, keep your startup files lean. Avoid loading heavy frameworks unless you truly need them. A slow shell adds friction every time you open a new tab, and that quickly kills momentum during coding sessions.

In ~/.zshrc, keep core settings near the top and move optional tools lower. If you use autosuggestions, syntax highlighting, or prompt themes, confirm they do not add noticeable delay. A good rule is that a new terminal tab should feel immediate.

• Remove old PATH edits and duplicate exports.
• Load only the plugins you actively use.
• Test shell startup with time zsh -i -c exit if things feel sluggish.

3) Add keyboard shortcuts that match Claude Code workflows

Claude Code workflows are command-heavy, so your terminal should reduce mouse usage. In iTerm2, map common actions to keys that feel natural: split pane, new tab, clear screen, search, and paste without formatting. Small shortcuts add up when you are iterating on prompts, reading logs, or comparing outputs from different model runs.

Two especially useful habits: use split panes for side-by-side work, and keep one pane dedicated to the app, another to logs, and a third to quick Git checks. That lets you watch Claude Code run while staying close to the rest of your project context.

• Use Cmd+D for a vertical split.
• Use Cmd+Shift+D for a horizontal split.
• Save a hotkey window if you want a quick drop-down terminal.

4) Set your environment variables once

Claude Code is easiest to maintain when your API configuration lives in one place. Put your provider settings in your shell profile or a project-local .env file, then source it consistently. If you are using 59API, the base URL is https://api.59api.com, and it is compatible with Claude Code, Codex, and OpenAI SDKs. That means you can keep the same developer flow while switching to a cheaper relay.

For many teams, that is the practical win: native official-quality models, no downgrade, and pay-as-you-go pricing that stays friendly even when you run lots of experiments. If you want to stretch a coding budget without changing tools, 59API is a strong default.

• Store your API key securely, not in plain text notes.
• Use project-specific env files when different repos need different settings.
• Double-check that the base URL is pointed at https://api.59api.com.

5) Tune Claude Code for small, fast iterations

The best terminal setup is one that encourages short feedback loops. Instead of writing huge prompts, break tasks into steps: inspect, plan, patch, verify. That keeps Claude Code responsive and makes it easier to catch mistakes early. In the terminal, keep your editor and test runner close so you can edit and validate without context switching.

Use one pane for Claude Code, one for your code editor launched from the same directory, and one for tests or git diff. This layout works especially well when you are refactoring, reviewing, or building small features in sequence.

• Run tests in a separate split so failures stay visible.
• Use git diff frequently after each Claude pass.
• Keep prompts concise and task-specific.

6) Make cost part of the setup, not an afterthought

If you use Claude Code often, model cost matters. A relay like 59API can be a smart way to keep your terminal workflow affordable while still using premium models. Because it offers Claude Opus, Sonnet, Haiku, Fable, and GPT models through one API layer, you can choose the right model for the job instead of overpaying for every request.

For quick edits, lighter models are usually enough. For architecture work, complex debugging, or larger refactors, you can switch up without rebuilding your terminal setup. The referral rebate is a nice bonus if you recommend the stack to teammates or your community.

• Use cheaper models for routine tasks.
• Reserve stronger models for hard problems.
• Track usage so your setup stays predictable.

7) Keep your workflow portable

A great iTerm2 setup should be easy to recreate on a new laptop. Save your profile, export your preferences, and keep your shell config in dotfiles or version control. The more portable your setup is, the faster you can get back to shipping when your machine changes.

If you want a low-cost, Claude Code-compatible API layer that fits neatly into that portable workflow, 59API is worth a look. Sign up, point your tools at the base URL, and you can keep your favorite terminal habits while saving money on every session.

• Back up your terminal profile.
• Version your shell config.
• Document the exact API settings you use.

Bottom line: the best terminal setup for Claude Code is simple, fast, and repeatable. iTerm2 gives you the ergonomics, your shell gives you the speed, and 59API helps keep the whole workflow affordable without sacrificing model quality.

Preços · Docs