Applications

Acquarello

Turn ideas into images. From word to image, or from one image to another — always in your style.

On this page

    Acquarello is an AI image generation application, built in Streamlit, that puts the power of generative models in anyone's hands — no code, no hassle. Describe what you imagine and get an image; or upload an existing image and reinvent it. In both cases, you define the visual style right in the sidebar and watch the result take on your look.

    #✨ Highlights

    • 🖋️ Text → Image — describe a scene and let the AI create it.
    • 🖼️ Image → Image — start from an image and turn it into something new.
    • 🎚️ Style under control — choose and adjust the style from the sidebar.
    • ⚡ Simple to use — intuitive interface, designed for any user.
    TitleLink
    💻 Source code (GitHub)Go to the repository
    ▶️ How to run the projectInstallation guide
    ⌨️ Terminal commandstreamlit run web_app.py
    🌐 Try it now (Streamlit)http://localhost:8501/acquarello

    #Context

    Generative image models are among the most impressive advances in recent AI — but, in practice, they still intimidate non-technical people: they require configuration, complex prompts and scattered tools. Acquarello was born to solve this, bringing together text generation, image generation and style control in a single accessible interface. The proposal is to bring anyone closer to visual creation with AI — from the curious to the creative — showing that this technology can be powerful and simple at the same time.

    #Use cases

    #1. Generation from text (text-to-image)

    #A sentence becomes an image. Ideal for creating concepts, illustrations and scenes from scratch

    #"Gere uma imagem aquarela com cores bem vibrantes e coloridas de duas crianças brincando com uma pipa na calçada"

    #"Retrato de uma senhora idosa sorrindo, com rugas marcadas e olhar expressivo"

    #"Rua de vila japonesa à noite com lanternas acesas e chuva leve"

    #"Uma família reunida na sala assistindo a televisão nos anos 1950"

    #2. Generation from image (image-to-image)

    #One image turns into another. Perfect for reinterpreting, remixing or refining existing visuals.

    #Original image

    #Image with Anime style applied

    #Original image

    #Image with Cartoon style applied

    #Original image

    #Image with Rabisco style applied

    Source: src/web_applications/applications/acquarello.py

    Esc
    ↑↓navigate Enteropen