"Changing a pose" splits into two genuinely different jobs. If you have a specific photo and want to adjust the pose in it, that's editing — Adobe Firefly (via Photoshop's Generative Fill) and RunwayML both do real, sourced versions of this. If you want a new photo of yourself in a different pose, in a setting you describe, that's generation — which is what Pose AI's Image Studio does, from a single selfie, with no existing photo required.
Both are legitimate answers to "how do I change a pose," they just start from different places — an existing image versus a blank canvas and your identity.
- Pose editing (Adobe Firefly, RunwayML): you start with a specific photo or video you already have, and the tool modifies part of it — including, in Firefly's case, the subject's pose — while keeping the rest of the image intact.
- Pose generation (Pose AI, via Image Studio and Nano Banana 2): you start with one selfie and a description, and the tool generates a brand-new, identity-locked photo of you in the pose, outfit, and setting you asked for — nothing existing gets modified.
- Pose AI does not edit an existing photo's pose. If that's specifically what you need, Adobe Firefly's Generative Fill is the better-fit tool.
- Firefly Image 5's instruction-based Generative Fill in Photoshop explicitly supports describing a pose change in natural language while preserving the rest of the photo; RunwayML's Act-Two model drives pose-and-lip-sync animation onto a character from a reference performance, and Edit Studio (Aleph 2.0) transforms existing footage by prompt.
Key terms
Pose editing means modifying the pose within a photo or video you already have, while the rest of the image — background, lighting, other details — stays as it was. It's a targeted change to an existing asset, typically done with a selection or a reference performance driving the adjustment.
Pose generation means creating a new photo from a text or reference-image prompt and, for identity-locked tools like Pose, a single uploaded selfie — there's no existing photo being modified, because the output is generated from scratch in the pose, outfit, and setting you describe.
How Pose AI's Image Studio actually works
Image Studio, powered by Nano Banana 2, reads your face from a single uploaded selfie and generates a new photo in whatever pose, setting, and style you describe — standing, seated, in motion, close-up or full-body. It is not opening a photo you already have and adjusting it; it's building a new one from your identity and your description, with no lengthy training step.
Generate a new pose from one selfie on the AI headshots page.
See the full range of poses and settings across Pose AI's style library.
True pose-editing tools
Adobe Firefly, via Photoshop's Generative Fill, is the more direct match if you specifically need to adjust the pose within a photo you already have. Firefly Image 5's instruction-based editing lets you describe a pose change in natural language — brush over the area, describe the adjustment, and Photoshop regenerates it while keeping the rest of your original pixels intact.
RunwayML approaches it from the video side: the Act-Two model drives pose-and-lip-sync-accurate character animation onto a subject from a reference performance, and Edit Studio (built on Aleph 2.0) lets you transform existing footage by prompt — restyling a shot, adjusting lighting, adding or removing elements. Both are real editing tools working on footage or images you already have, which is a different starting point from Pose's one-selfie generation.