Claude Code Keyboard Shortcuts & Tips: The 2026 Cheat Sheet
Claude Code is fast. But it is even faster when you know the shortcuts. Whether you are a first-time user or someone who has been shipping code with Claude for months, having every essential shortcut committed to muscle memory makes a real difference in your day-to-day output. This cheat sheet covers every essential Claude Code shortcut, slash command, and power-user tip -- plus the Beam shortcuts that make managing multiple sessions effortless. Bookmark this one.
Claude Code Essential Shortcuts
These are the commands and shortcuts you will use constantly while working with Claude Code. They break down into two groups: commands for launching and quitting sessions, and slash commands you run inside a live session.
claude
Start a new Claude Code session in your current directory
claude -r
Resume your last session (keeps full conversation context)
claude "prompt here"
Start a session with a specific prompt inline
/exit
Exit Claude Code and return to your shell
/clear
Clear conversation history and start fresh without exiting
/compact
Compress the conversation to save context window space
/cost
Show token usage and cost for the current session
/doctor
Diagnose configuration issues and check your setup
/help
Show all available slash commands
/init
Create a CLAUDE.md project memory file in your repo
/model
Switch to a different model mid-session
Power User Tips
Knowing the shortcuts is only half the battle. How you use Claude Code in practice determines whether it saves you minutes or hours every day. These are the habits that separate casual users from people who genuinely ship faster with AI.
Resume Sessions
claude -r picks up exactly where you left off. Do not lose context by starting fresh every time -- resume and keep building on what Claude already knows about your codebase.
Pipe Input Directly
Use cat error.log | claude "explain this error" to pipe files or command output straight into Claude. No copy-paste needed.
Project Memory
/init creates a CLAUDE.md file in your project root. Claude reads this automatically every session, giving it persistent context about your stack, conventions, and preferences.
Work Across Multiple Files
Claude can read and modify multiple files in a single operation. Do not limit yourself to one file at a time -- describe the full scope of the change and let Claude handle it.
Let Claude Run Commands
Claude can execute shell commands for you -- building, testing, installing packages, running migrations. Tell it what you need done and it will handle the terminal side.
Use /compact Proactively
When conversations get long, /compact preserves the important context while freeing up the context window. Run it before you hit the limit, not after.
Write Better Prompts
Specificity is everything. "Refactor the auth module to use JWT with refresh tokens, update the middleware, and add tests" will get you dramatically better results than "improve the auth." The more precisely you describe the outcome you want, the fewer iterations you will need. Include file paths, function names, and expected behavior when you can.
Beam Shortcuts for Claude Code Sessions
Claude Code runs in the terminal, which means your terminal app matters. Beam is built specifically for managing multiple terminal sessions. Here are the shortcuts that make the biggest difference when you are juggling several Claude Code instances across projects.
The Speed Workflow: Combining Both
The real productivity gain comes from combining Claude Code's commands with Beam's session management. Here is what a keyboard-only workflow looks like in practice -- from opening Beam to shipping code, without touching the mouse once.
claude -r. You are back in the conversation with full context.Customizing Claude Code
One of Claude Code's most powerful features is project memory. Claude automatically reads a file called CLAUDE.md from your project root at the start of every session. This gives it persistent context about your codebase, your conventions, and the way you like things done -- without you repeating yourself.
Use it to define your stack, set coding conventions, and provide project-specific instructions that Claude should follow in every interaction. The more detailed your CLAUDE.md, the less time you spend correcting Claude's assumptions.
If you are using Beam, you do not even need to create this file manually. Beam's "Install Project Memory" toolbar button generates a CLAUDE.md tailored to your project and drops it into your working directory. It is one click to give Claude persistent context about what you are building.
Master Your AI Coding Workflow
Download Beam for the terminal shortcuts that make Claude Code even more powerful. Workspaces, tabs, splits, and a Quick Switcher -- all keyboard-driven.
Download Beam for macOSSummary
Here are the essential shortcuts to commit to memory. Once these are second nature, you will never go back to clicking around terminals or restarting conversations from scratch.
Claude Code Essentials
claude -r-- resume your last session with full context/compact-- compress conversation to save the context window/clear-- start a fresh conversation without exiting/init-- create a CLAUDE.md for project memory/cost-- check your token usage and cost- Ctrl+C -- interrupt a response
- Ctrl+D -- exit Claude Code
Beam Essentials
- ⌘N -- new workspace per project
- ⌘T -- new tab
- ⌘⌥⌃T -- split pane
- ⌘P -- Quick Switcher to find any session
- ⌘S -- save your layout
- ⌘Z -- undo close
- ⌘1–⌘9 -- jump to tab by number
Happy coding.