Best Practices

Atom keeps you moving fast. These quick habits make sure the results stay dependable.

Protect your work

Save before big automations so Undo always has a safe place to land.
Duplicating a comp for experiments lets you compare versions without stress.
Keep a simple backup naming rhythm (Project_v01.aep, Project_v02.aep).

Stay clear and organised

Use descriptive layer names. Atom can only hit what it can identify.
Stick to a consistent stack (controls on top, hero layers in the middle, utilities below) to keep prompts predictable.
Tighten the scope of each request (“selected layers”, “Scene 03 only”) for faster, cleaner runs.

Iterate with intent

Build complex ideas in small steps and confirm after each reply.

When working with expressions, ask Atom to outline dependencies before renaming or deleting layers.

Check History to reuse great prompts and evolve them into a personal library.

Tune the chat

Enable Settings → Features → Auto-scroll chat to keep long responses pinned to the latest update.
Turn on Notification sound and choose a completion sound if you want an alert after longer tasks.
Use Clear chat history when you want to remove all saved conversations.

Watch the meters

Trial days count down in the top bar. Upgrade before the trial ends.
Monitor your AI usage through your provider dashboard.

Capture what works

- Turn proven patterns into [skills](/docs/reference/skills) so Atom can reapply them across projects. - Note repeatable prompts for teammates and keep conventions documented.

Ask Atom to create a skill after you’ve built something worth reusing: “Create a skill from this animation style.”