ChatGPT and Codex Launch New Image Editing Workflow with Canvas and Pointing
OpenAIDevs · x · 2026-07-31
OpenAI has introduced a new image editing workflow for ImageGen within ChatGPT and Codex. The update features a new lightbox and canvas, making it significantly easier to explore and refine visuals.
Instead of writing long, complex prompts to make edits, users can now interact directly with the generated image. The new workflow allows users to point at specific characters, erase unwanted elements, and add comments exactly where changes are needed, greatly streamlining the creative process.
More from Multimodal
- Tutorial: Local Image Classification in SwiftUI Using Core AI and CLIP — amos_gyamfi · 2026-07-31
- AI Creative Event Preview: Tackling GenAI Flaws and Music Model Demos — stuffyokodraws · 2026-07-31
- Google Vids Integrates Gemini Omni to Generate Personal AI Avatars from a Selfie — aziz4ai · 2026-07-31
- Indie Game Showcases AAA Procedural Character Controller — majidmanzarpour · 2026-07-31
- Creating Cinematic Escape Videos with ChatGPT and Hedra — aziz4ai · 2026-07-31
- Microsoft Open-Sources TRELLIS: Turns Any Image into 3D Assets in 3 Seconds — OwariDa · 2026-07-31