Use Cases
LearnFork for Developers
The problem with linear chat
You're debugging a React component and the AI suggests using useCallback. You're not sure when useCallback actually helps, so you ask. Then you wonder about useMemo too. Now your debugging chat is 40 messages deep and half of them are about hooks in general, not your bug.
Developers need to explore multiple hypotheses, check documentation mid-task, and compare approaches — all without losing their original thread.
The LearnFork way
Your main conversation is the task. Branches are explorations:
- Vertical (keep chatting) — work through your main task step by step. Stay focused.
- Horizontal (branch) — when you need to understand a library, check an API, explore an alternative approach, or investigate an error — branch. Get your answer, then come back.
Step by step: debugging a bug
- Start a conversation describing your bug: paste the error, the relevant code, and what you expected
- The AI suggests a fix involving a pattern you're not sure about. Branch from that response.
- In the branch, ask: "Explain how this pattern works and when to use it"
- Back in the main thread, the AI suggests an alternative approach. Branch from that response too.
- Now you have two branches — two different approaches you can compare side by side on the canvas
- Continue in whichever branch has the better solution
Learning a new framework
Start a root conversation: "Walk me through building a REST API with [framework]." Branch whenever the tutorial mentions something you need to understand better — middleware, ORMs, authentication patterns. Each branch becomes a mini-lesson with full context of what you're building.
Tips
- Branch from a code suggestion to ask "What if I used X instead?" — compare approaches on the canvas
- Use "Go deeper" on error explanations to understand root causes, not just fixes
- Pin code snippets you want to reference later
- Select unfamiliar API names in responses and use "Google" to check the official docs
- Keep your main thread clean — branch for every tangent, no matter how small