Tips and Tricks for Happy Coder

Happy Coder gives you control over AI coding agents from anywhere. This guide covers patterns and workflows that help you get the most out of mobile coding sessions.

1. Set up for success

a. Connect everything, everywhere

The biggest workflow improvement you can make is connecting all your devices and browsers to the same account. Don’t just connect your main laptop. Connect every computer that can run code.

Connect your:

Development machines: Laptop, desktop, Mac mini in the closet
Cloud resources: VPS, dedicated servers, cloud instances.
Edge devices: Raspberry Pi, router running OpenWRT, home server
Your browsers: Happy Coder isn’t just a mobile app, you can use the web version designed for larger screens. You’re never more than a browser tab away from checking on your AI agents.

Once everything is connected, all your sessions sync everywhere. You can start a session from anywhere and pick it up from anywhere else. Start coding on your desktop, check progress from your phone at lunch, then review the PR from your tablet on the couch. Every device shows the same sessions, the same progress, the same capability.

Diagram needed: Network diagram showing multiple devices (laptop, phone, tablet, VPS, Pi) all connected to the same Happy Coder account, with arrows showing session handoff between devices.

b. Turn on experimental features

The experimental features flag in Settings is where all the active development happens. Turn it on immediately to access the latest capabilities. New features ship here first before graduating to stable.

Screenshot needed: Settings screen showing the experimental features toggle enabled, with a callout highlighting this as the first thing new users should enable.

c. Switch agents freely

You’re not locked into just using one of Claude Code or Codex. Try different agents for different tasks:

Keep working with projects already set up with Claude Code
Try using Codex on the same project to see how it compares
Try other agents as they become available

The input field lets you switch between agents on the fly. Experiment to find what works best for each type of task.

Screenshot needed: Input field showing the agent selector dropdown or toggle, demonstrating how to switch between different AI coding agents mid-session.

2. Remember: Your terminal still works

Think of mobile as optional control, not replacement

Happy Coder enhances your existing workflow - it doesn’t replace it. You can still start Claude Code sessions with happy and Codex sessions with happy codex directly in your terminal on your computer. The mobile app gives you optionality, not obligation.

Think of it as seamless handoff: Work in your terminal when you’re at your desk, switch to your phone when you step away. The session continues exactly where you left off. Zero workflow disruption.

Happy Coder extends your existing workflow—it doesn’t replace it. You can still use claude or codex directly in your terminal. The mobile app gives you optionality: check on long-running tasks, approve changes, or start new work when inspiration strikes.

The seamless handoff is the key feature. There’s zero impact on your existing workflow. Work normally at your desk, then continue from anywhere when needed.

b. Start multiple parallel sessions

Run several AI agents simultaneously on different machines or even the same machine. Have one session refactoring your frontend while another writes tests for your backend. Switch between them instantly from your phone.

This works especially well when you have different types of compute available:

Use your desktop’s GPU for ML tasks
Use your VPS for deployment and testing
Use your laptop for local development

Screenshot needed: Mobile app showing multiple active sessions in a list or tab view, with clear labels showing which machine and which project each session is working on.

3. Set up git worktrees ahead of time

Happy Coder doesn’t manage git worktrees for you (yet), but with a little setup, you can make experimentation seamless:


# Create a few worktrees for experiments
git worktree add ../project-experiment-1 -b experiment-1
git worktree add ../project-experiment-2 -b experiment-2
git worktree add ../project-experiment-3 -b experiment-3

Now from your phone, you can start new sessions in these directories whenever you want to try something. Each worktree is isolated, so you can run multiple experiments in parallel without conflicts.

Screenshot needed: Mobile interface showing a file browser with multiple git worktree directories, demonstrating how to start a new session in a specific worktree from your phone.

4. Build new habits

a. Practice the “instant idea” muscle

The ability to tell AI to make a change from anywhere is genuinely new. Build this muscle memory by practicing, even with small changes. Walking the dog and think of a better variable name? Pull out your phone and make it happen.

Start small:

Fix that typo you just remembered
Add a comment you’ve been meaning to write
Try that refactoring idea you had in the shower

The goal is making “have idea → implement idea” as frictionless as checking your messages.

b. Do a personal inventory

Ask yourself: “What’s holding me back from just trying ideas?”

Common blockers and solutions:

Fear of breaking things: Set up better git workflows, use branches liberally
Unclear how to revert: Learn git reset/revert commands, make more commits
Worried about wasted time: Remember that exploring ideas quickly is valuable
Perfectionism: Ship experimental branches, iterate in public

Until Happy Coder handles more of these automatically, identify your personal blockers and address them directly.

Diagram needed: Simple flowchart showing the ideal “idea → implement → test → ship/revert” cycle, with Happy Coder removing friction at each step.

5. Quick workflow patterns

a. The commute pattern

Start a big refactoring task before leaving work
Check progress on your commute
Approve or redirect from your phone
Arrive home to completed work

b. The inspiration pattern

Have an idea while away from computer
Open Happy Coder on your phone
Describe what you want in natural language
Let the agent work while you continue your day

c. The parallel exploration pattern

Start 3-4 different approaches to a problem
Let them run on different machines/sessions
Check results on your phone
Pick the best approach, discard the rest

Key takeaway

Happy Coder’s power is to extend your desktop workflow and tools everywhere. Connect all your compute, build the muscle memory of describing problems and reviewing implementation, and use the freedom to explore more ideas faster. The best code often comes from trying things you wouldn’t have bothered with before.