Launching today
Canvasory
Turn source material into art-directed visual presentations
2 followers
Turn source material into art-directed visual presentations
2 followers
Canvasory turns source material into image-based presentations. Its AI art director plans each page’s focus, imagery, relationships, layout, and reading order. Review the split and shared direction, then generate, preview, and export high-res PNGs as a ZIP.
Hey Product Hunt 👋
I'm Sander, working with @leven_xie on Canvasory.
Codex already makes beautiful images. The problem I kept hitting was different: I had content I wanted to turn into a visual deck, but every slide needed its own visual brief — the focus, the imagery, the layout.
All from scratch, every time.
A deck about a frog's life cycle changed how I thought about this.
When we generated images directly from the content, each slide looked fine on its own.
But when we added an art direction step — planning the visual story before generating anything — the stages of metamorphosis connected into one continuous scene.
Same content, same model. Way clearer story.
That's the core of Canvasory: an AI art direction layer that sits before image generation.
Before rendering anything, it plans what each slide should communicate — its focus, imagery, relationships, layout, and reading order.
You bring the content and a creative direction.
It handles the per-page visual decisions you'd otherwise make yourself.
How it works:
→ Add your material
→ Review the page outline
→ Choose Auto or set one shared visual direction
→ Canvasory plans each page, renders the deck, and gives you a preview + ordered high-res PNGs
One honest trade-off: these are image-based slides, not editable PowerPoint objects.
We're building for illustrated explanations, comparisons, and visual stories — not for decks that need constant text or chart edits.
This release runs as a Codex Skill on macOS ARM64 and Linux x86_64, so the workflow stays inside Codex rather than adding another presentation editor.
What's something you know well but find hard to explain visually? Curious what you'd try first.