Stop Typing Prompts and Start Directing Scenes: ArtCraft Gives Artists Real Control

You know the frustration: you type a beautifully detailed prompt into an image generator, hit generate, and get something that's almost right—but the character's hand is in the wrong place, the background doesn't match the other shots in your sequence, and you can't seem to get two frames to feel like they're in the same room. Text-to-image is powerful, but it's also a bit like shouting directions at a blindfolded painter. What if you could actually step into the scene, move things around, and compose the shot yourself? That's exactly the problem ArtCraft is built to solve.

ArtCraft is an open-source IDE for interactive AI image and video creation. The project's pitch is bold: turn prompting into crafting, where your ideas become tangible, manipulable elements rather than one-shot text gambles. They call it "Adobe Photoshop for everyone," and they're open-sourcing the whole thing.

What It Does

At its core, ArtCraft is a visual environment where you don't just describe what you want—you build it. The README makes a strong case that text-to-image alone isn't enough for serious artists. You need to know what your image will look like before you generate it, and you need consistency and repeatability across multiple outputs.

The project delivers this through a suite of compositing and scene-blocking tools. You can import images, turn them into 3D meshes, place virtual actors into physical environments, pose characters dynamically, and use 3D asset kits to block out entire scenes with precise angles and depth layering. The workflow is about assembling a scene from tangible pieces—images, meshes, layers—and then letting the AI render the final result.

The README is light on technical implementation details, but the feature set points to a desktop application that integrates AI generation with a canvas-based editing environment. Think layer-based editing (like Photoshop) combined with 3D scene layout (like a game engine editor) and AI image generation under the hood. It's a heavyweight tool for a heavyweight problem: making AI art reproducible and controllable.

Why It's Cool

What makes ArtCraft interesting isn't any single feature—it's the underlying philosophy. Most AI art tools treat the prompt as the primary interface. ArtCraft treats the scene as the interface. Here's what stands out:

• Image to Location is a killer feature for storytellers. The README describes placing virtual actors into physical environments to establish single-location consistency. That means you can film multiple shots within a room and things won't mysteriously vanish or change between frames. For anyone trying to create a coherent narrative with AI video, that's enormous.

• 3D Image Compositing bridges the gap between static images and dynamic scenes. You can take backdrops, props, and foreground elements, give them depth, and blend them naturally. The README notes that just a couple of images usually leads to great compositions—so you're not spending hours on asset creation.

• Image to 3D Mesh solves a real pain point. Laying out complicated objects or blocking complex scenes with text is nearly impossible. By converting images to 3D meshes, you can position elements exactingly and intentionally. This is the difference between saying "a vase on a table" and actually placing the vase where you want it, at the angle you want, with the lighting you want.

• Character Posing lets you dynamically pose characters before you generate. You can achieve precise character, scene, and camera blocking before you call "action." It's directing, not hoping.

• Scene Blocking with Kit Bashing is a technique borrowed from visual effects. You use 3D asset kits to block out your scene—getting the correct angles, object positions, and rich depth layering—which you simply can't achieve with text prompting alone.

The common thread is control. ArtCraft treats AI generation as the final render step in a pipeline where you, the artist, make all the important decisions about composition, placement, and consistency. That's a fundamentally different approach from prompt-and-pray.

How to Try It

ArtCraft is open source, and the team is clearly building a community around it. You can find the repository on GitHub at github.com/storytold/artcraft.

The README doesn't include quick install commands—this appears to be a project that's still evolving its onboarding documentation. Your best bet is to head to the repository and check the README directly for the latest setup instructions. If you want to see the features in action before you commit to a local install, the README is full of embedded demo videos showing each workflow.

You should also join the community. The project maintains an active presence on Discord where you can ask questions, share what you're building, and get help from other users. There's also a YouTube channel with subscriber content, and they're active on X (formerly Twitter) as @get_artcraft.

Final Thoughts

ArtCraft is for anyone who's hit the ceiling of text-only AI generation. If you're a filmmaker storyboarding scenes, a game developer blocking environments, or an illustrator who needs consistent characters across multiple images, this tool speaks your language. It's ambitious—calling yourself "Photoshop for everyone" sets a high bar—but the feature set suggests the team is serious about giving artists genuine control over the AI creative process.

The project is open source, which means it's not just a tool you use; it's a tool you can shape. Whether ArtCraft becomes the standard for interactive AI creation remains to be seen, but it's clearly pointing in a direction the industry needs: away from blind prompting and toward deliberate, repeatable, artistic craft.

Follow @githubprojects for more developer tools and open source projects.