Your first win with Chat.
Brief a move in plain language, watch Altar build the keyframes, and read exactly what it changed.
Before you start
Altar Chat runs on your own AI provider key. Altar never charges you for inference. Add a key for Anthropic, OpenAI, or Google in Settings, or point Chat at a local model, then open the comp you want to work on.
Describe a move
Type the move the way you would say it to a colleague:
Bring the title up over twelve frames and settle it with a soft overshoot.
Altar reads the comp first, works out which layer and which property you mean, then puts real keyframes on the real property.
You are not writing expressions or clicking through panels. You describe the result; Altar Chat builds the keyframes and places them on the timeline.
Read the receipts
Every move comes back with a receipt - what changed, on which layer, on which property. Read the receipt, not a wall of jargon. If it is not what you meant, say so, and Altar adjusts.
- Describe the move in plain language
- Altar applies it through a validated operation
- The receipt tells you exactly what landed
Keeping the taste
Chat is fast hands that respect your timeline. The composition and the judgement stay yours; the repetitive part is what gets handed over.
Next steps
- Prefer to do it by hand? See your first win with Altar.
- Want the technical picture of what Chat can see? Read How Altar works.
Replace fonts
Try “Change every font in this composition to Arial Bold,” using a font installed on your machine. Chat inspects the composition first. If scope or weight is missing, it asks before applying the change. You can preserve existing weights, choose one face, include child compositions, or limit the change to selected text.
The font operations guide explains supported text, style substitutions and recovery.