⚒️ AI Creator
Nano Banana Pro
This is 'nano-banana', Google's most powerful image generation model. Create new images from text, edit photos with simple commands, or merge multiple pictures into one—simply describe what you want to create or change. **Image Uploads:** You can also upload an image directly. If an image is uploaded, it will be automatically referenced as "this image" in your subsequent commands. Usage examples: Create a new image: > <@<intern_id>> Generate an image of a cat sitting on a laptop. Edit an existing image: > <@<intern_id>> Edit this image: remove the background and make it transparent. Merge multiple images: > <@<intern_id>> Merge these two images: place the moon behind the mountain.
All-in-one AI image
This all-in-one image generation mode automatically selects the best model for your task, drawing from: • Flux: For high-fidelity rendering. • ChatGPT: For creative visual capabilities. • Dream 4.0: For extraordinary imagination. • Nano Banana: For high-speed generation. No need to choose a model—just describe your idea to get the highest-quality visual work. **Image Uploads:** You can also upload an image directly. If an image is uploaded, it will be automatically referenced as "this image" in your subsequent commands. Usage examples: > <@<intern_id>> I need a sci-fi concept art of a futuristic city, with flying cars and holographic billboards. > <@<intern_id>> Design a cover for my podcast "Late Night Talks" in a minimalist, artistic style.
Mini Web Page & Game
Build web page, games, and interactive elements with your team and friends—no code required. Simply describe your idea in plain language to generate the complete HTML. Each creation also comes with explanations of the UI/UX design thinking behind it. Usage examples: > <@<intern_id>> Make a simple game of Tic-Tac-Toe. > <@<intern_id>> Design a button that lights up when you hover on it. > <@<intern_id>> Build a login page for my new app.
Veo3 Video
**# IMPORTANT: This model is expensive to operate. `fast` mode costs a minimum of 160 TUs and `normal` mode costs 400 TUs per generation. Please use it judiciously.** ** I am your fully autonomous AI Video Production Assistant, powered by the Veo3 engine. My purpose is to bring your creative visions to life, transforming your ideas—whether from a detailed text description or a static image—into a high-quality, cinematic video clip of up to 8 seconds. I autonomously manage the entire production process, and you can always: * **Animate an image** by providing a picture and describing the desired motion. * **Check the progress** of a video currently being generated. * **Recall previous creations** by asking me to show you your recent videos. ### Usage examples: **For Creating a New Video (Text-to-Video):** > <@<intern_id>> Create an 8-second video of a golden retriever puppy playing in a field of flowers, cinematic slow motion. > <@<intern_id>> I need a short opening shot: a futuristic city street at night, with rain reflecting neon lights and flying cars speeding by. **For Animating an Image (Image-to-Video):** > `[User uploads a picture of a majestic lion]` > <@<intern_id>> Make its mane blow gently in the wind, slow motion. **For Managing Your Videos:** > <@<intern_id>> What's the status on the futuristic city video? > <@<intern_id>> Show me my last few videos.
像素风滤镜
# 👾 Pixel Art Portrait Filter Welcome to the "Pixel Art Portrait Filter" workflow, your dedicated **automated retro-style converter**. This function is designed as a "one-click" experience, focused on rapidly transforming your uploaded **portrait photos** (especially female portraits) into a highly consistent **Pixel Art** style. It does not rely on complex text instructions; instead, it uses a fixed library of advanced parameters to ensure the best artistic effect with every conversion. ## ✨ Core Features & Aesthetic Style * **One-Click Filter:** This is a pure **Image-to-Image (I2I)** workflow. You do not need to write any prompts. Simply upload your photo, and the Agent will automatically apply all preset parameters. * **Fixed Aesthetic Style `[FILTER_PIXEL_V1]`:** * **Core Style:** Utilizes a specific Lora (`mPixel_v20PixelStyle`) and a Denoising strength of 0.4 to apply a stable, high-quality pixelation art effect while preserving the structure of the original portrait. * **Portrait Optimization:** The default prompt (`1girl`) is optimized for female portraits, ensuring the best pixelation effect on faces and features. * **Auto-Hires Processing (Auto-Hires):** The workflow includes a built-in **1.3x high-definition upscale** (`Hires upscale`), using the `R-ESRGAN 4x+ Anime6B` upscaler to ensure the final pixel artwork remains sharp and refined even after enlargement. * **Locked Parameters:** All generation parameters (e.g., Steps: 25, Sampler: DPM++ SDE Karras, CFG scale: 7) are strictly fixed to guarantee **absolute style consistency** and predictable output results. ## 🚀 Getting Started * **Upload to Convert:** * This workflow **requires only one input**: your portrait photo. * Upload your photo, and the Agent will automatically retrieve the `[FILTER_PIXEL_V1]` library and begin the stylization process. > [Upload your portrait photo.jpg]