Image Studio

The Image Studio is River's image generation and editing surface. Generate brand-consistent visuals, edit existing images with AI, and produce assets without leaving the workspace.

What you can do#

• Generate — text-to-image with multiple model tiers (Standard, HD, Uncensored)
• Edit — upload or open any Image entity and edit with natural-language instructions
• Variations — produce N variants of an image for review
• Remix — feed an image as a reference for new generations
• Background remove / replace — one-click
• Upscale — turn a 512×512 generation into a print-ready asset
• Inpainting — mask and regenerate a region

Models#

Image Studio supports multiple providers:

For self-hosted River + local image models, see Local Models.

Output as entities#

Every generation produces an Image entity in your Dataspace, with the prompt, parameters, and source images all stored as metadata. You can:

• Search past generations by prompt
• Re-run with one tweaked parameter
• Link the image to a project, a doc, or a launch
• Pass the image to other tools (a slide deck, a social post Workflow, a Custom App)

AI use patterns#

• Brand assets at scale — generate hero images for every blog post in one batch.
• Mockups — turn a Doc description into an illustrative image inline.
• Social content — pair Image Studio with a Workflow that generates and posts.
• Variations — generate 8 variants of a hero, pick the best, edit it.

Cost#

Cloud image generation has per-image costs that vary by provider and quality tier. See Pricing for current rates.

Local generation is free after the hardware. A consumer GPU can run Stable Diffusion variants at quality high enough for most use cases.

Tips#

• Save winning prompts. Reuse them; they compound over time.
• Generate in batches of 4–8 variants. Pick, then refine.
• For brand consistency, build a Skill that prepends your style guide to every prompt.