Q: Requests that are really important for user experience
Could you consider adding the following features:
1. Reference option (like Runway): Would like an option where we can Upload a reference image, then separately upload my avatar model picture. Include a prompt box for atmosphere, background, clothing, and aesthetics. The system would recreate the same pose and composition but replace the original person with my model.
2. Relight feature (like Photoshop): Match lighting when adding a person into a background they weren’t originally in.
3. Favorites option in prompts: Save repetitive prompts for quick reuse.
4. Simple face swap tool for one-off images without training.
5. Built-in negative prompts such as “no morphing” and “no disfigured limbs.”
Thanks for considering!

Creative_Bot_PhotoGPTAI
Aug 12, 2025A: Thanks for the thoughtful and well-structured suggestions — these are exactly the kind of insights that help shape our roadmap. Here’s where we stand on each:
- Reference image: This feature is already available for deriving style and composition. You can also control the Style Strength setting to adjust how much influence the reference image has on the output.
- Relighting: This isn’t currently on our roadmap, but we’ll take it back to the team for discussion. Feel free to join our Discord community (https://discord.gg/rENRJw439H) to share more context and get live updates.
- Favorites in prompts: Great idea, noted! We’ll keep you posted as we move in this direction.
- Face swap: Not in near-term plans, as we value realism and current face-swapping techniques aren’t accurate enough. That said, you can try Generate Similar and switch the model, or use the Reference Image feature to place a model into a specific style.
- Negative prompts: We agree negative prompting can be tricky. Instead, framing instructions in an affirmative way (e.g., “perfectly proportioned limbs”) tends to be more effective. We’ll discuss with the team about improving support for negative prompts.
Thanks again for sharing, it’s feedback like this that drives continuous improvement.