Daily Productivity Boosters for Codex

The difference between a 10-minute Codex task and a 3-minute one is rarely the model — it is the workflow around the model. Shell completions, file references, voice dictation, session resumption, and a few non-obvious keyboard shortcuts can cut your daily Codex time in half without changing how you think about prompts.

What You’ll Walk Away With

• 15+ specific techniques that individually save seconds but compound across a day
• CLI shortcuts and keyboard bindings you may not know exist
• Session management patterns that eliminate context re-entry
• Configuration tweaks that reduce friction on every Codex interaction

CLI Speed Boosters

Shell Completions

Install completions for instant tab-completion of all Codex commands and flags:

# Zsh (add to ~/.zshrc)
eval "$(codex completion zsh)"

# Bash (add to ~/.bashrc)
eval "$(codex completion bash)"

# Fish
codex completion fish | source

Now codex ex<TAB> expands to codex exec and flag names autocomplete.

File References with @

Type @ in the CLI or App composer to open a fuzzy file search over your workspace. Press Tab or Enter to drop the path into your message. This is faster than typing long file paths manually.

Quick Shell Commands with !

Prefix a line with ! to run a local shell command without leaving Codex:

!git log --oneline -5
!npm test -- --filter=auth
!cat src/routes/users.ts | head -20

Codex treats the output as context for your next prompt.

Session Resumption

Never re-explain context. Resume where you left off:

# Resume the most recent session
codex resume --last

# Resume from any directory
codex resume --last --all

# Resume a specific session
codex resume abc12345-...

Prompt: Quick continuation pattern

Instead of starting a new session with full context, resume and add:

codex resume --last "The approach we discussed works. Implement it and run tests."

This carries over the entire previous transcript.

Fork a Session

Want to explore an alternative approach without losing your current thread?

codex fork --last

This creates a new thread with the full transcript of your previous session, letting you branch the conversation.

App Speed Boosters

Keyboard Shortcuts

Shortcut	Action
Cmd + J	Toggle integrated terminal
Cmd + K	Command palette
Ctrl + M (hold)	Voice dictation
Ctrl + L	Clear terminal
Esc (2x)	Edit previous user message
Enter (while running)	Inject instructions into current turn
Tab (while running)	Queue a follow-up prompt for next turn

Voice Dictation

Hold Ctrl + M while the composer is visible and start talking. Codex transcribes your speech. Edit the transcription or hit send directly. This is faster than typing for complex, conversational prompts.

Image Drop

Drag and drop screenshots into the composer. Hold Shift while dropping to add as context rather than replacing the current prompt.

Configuration Tweaks

Clickable File Citations

Make file paths in Codex output clickable:

~/.codex/config.toml

file_opener = "cursor"  # or vscode, vscode-insiders, windsurf

Now /home/user/project/main.py:42 becomes a clickable link that opens the file at that line.

Prevent Sleep During Long Tasks

In the App settings, enable “Prevent sleep while running” so your machine stays awake during long Codex tasks and automations.

Profiles for Different Workflows

~/.codex/config.toml

[profiles.review]
model = "gpt-5.3-codex"
model_reasoning_effort = "high"
approval_policy = "never"

[profiles.quick]
model = "gpt-5.1-codex-mini"
approval_policy = "on-request"

Switch profiles on the fly: codex --profile review or codex --profile quick.

Disable Animations for Faster TUI

[tui]
animations = false
show_tooltips = false

Custom Notifications

Get notified when a task completes:

notify = ["terminal-notifier", "-title", "Codex", "-message"]

Or restrict TUI notifications to specific events:

[tui]
notifications = ["agent-turn-complete", "approval-requested"]

Prompt: Audit your Codex config for efficiency

“Read my ~/.codex/config.toml and suggest optimizations: are there unused MCP servers I should disable? Is my AGENTS.md too large? Am I missing any efficiency-boosting settings like shell_snapshot, file_opener, or profiles?”

Daily Workflow Template

1. Morning: Check the Triage inbox for overnight automation results
2. Planning: Use codex "Propose a plan for today's tickets" with your issue tracker MCP
3. Development: Run 2-3 parallel worktree threads for independent tasks
4. Review: Use /review before committing, then $pr-ready to prepare the PR
5. End of day: Archive completed threads, check /status for remaining limits

When This Breaks

• Shell completions not working: Ensure the eval line is after any compinit calls in your shell config. Restart your shell after adding it.
• Voice dictation inaccurate: Speak clearly and at a moderate pace. Edit the transcription before sending. It works best for conversational prompts rather than code-specific terminology.
• Profiles not loading: Check that the profile name matches exactly (case-sensitive). Use codex --profile <name> with the exact name from your config.
• Notifications not firing: On macOS, check System Settings > Notifications for both the Codex App and your terminal emulator.

What’s Next

• Prompt Engineering — Combine these speed techniques with better prompts
• Context Patterns — Manage context to keep sessions efficient
• Multi-Agent Workflows — Scale your daily workflow with parallel agents