Whether you're a total beginner, a digital art junkie, or a brand looking to dominate your niche with visual power, Touchy's AI image and video generator is your ultimate creative weapon. Built with simplicity at its core and flexibility at its peak, this tool is designed to make anyone look like a pro—no coding, no complicated setup, just plug in your vision and hit "Generate." From jaw-dropping illustrations to fluid, cinematic AI videos, you’ll be creating high-quality content in seconds, not hours. And we’re not talking basic filters or random results—you’ll wield deep control with full customization of every single part of your generations with every advanced tool at your disposal for any task, big or small. Need 4K resolution? We got it. Need to crank out twenty variants with different styles while sipping coffee? Done. Dial in every variable—sampler, steps, aspect ratio, even NSFW filters like a director on set. With hundreds of models, LoRAs, DoRAs, embeddings, and LyCORIS options added weekly, your creative arsenal never stops growing. Whether you're scaling your brand or unleashing your imagination for fun, this is the generator built to move fast, look sharp, and stay easy. You bring the idea and we’ll bring it to life, pixel by pixel.

The Workflow setting lets you choose how your image is created. If you're starting with just a written prompt, use Text to Image. If you're uploading a picture to use as inpiration, change or improve, use Image to Image. Face Fix fixes weird or messy faces and Hi-res Fix upscales and boosts image quality. Text to Image generates pictures from text. Face Fix fixes weird or messy faces while Hi-res Fix upscales and boosts image quality. Image to Image lets you start with a photo and change it using a prompt, with the same optional fixes available. When you choose an image-based option, an upload box will appear where you can drag or select a photo (PNG, JPG, WEBP, up to 15 MB).

Base models are the foundation of any generated image or video. When you choose a model, you’re selecting the core style, quality, and personality of the output. Each model is unique and brings a different vibe to the table. For example, SDXL focuses on creating high-quality, realistic visuals, perfect for lifelike scenes. Pony, on the other hand, is designed to generate stylized, cartoon-like images with sharp, clean lines and vibrant colors. Models like AnythingV5 strike a balance between realism and artistic interpretation, producing more detailed, diverse images. You can only pick one base model, and it will be the main framework for your generations. This decision is crucial because the base model impacts everything, from how characters look to the background details, setting the tone of the entire output.

Once you've selected your base model, you can enhance it with 9 additional resources like LoRAs, Embeddings, DoRAs, and Lycoris. These resources interact with the base model, helping to tweak or completely change certain elements of the image. LoRAs are mini add-ons that tweak the base model. They help adjust things like character details or overall style without changing everything. Embeddings teach the model specific concepts, like an artist's style or certain patterns, to give you a more customized result. DoRAs are like filters that change the overall vibe or theme of an image, like adding a sci-fi or vintage look. Lycoris refines details like lighting, texture, or small elements to make the image more polished and realistic. These resources allow you to fine-tune and specialize your base model, giving you even more creative control. They interact with the base model to either refine details or completely alter the style and output, making them essential for creating unique and personalized results.

Remixing lets you take an image you've already generated and build on it, whether you want to improve it, create something similar, or go in a completely new direction. When you click to remix an image, you're essentially using it as a foundation for a fresh creation. You can tweak your generation settings, change prompts, adjust styles, or even apply different effects. This opens up endless possibilities, allowing you to refine details, try new variations, or explore new themes without starting from scratch. It's an awesome way to keep iterating and experimenting with ideas, giving you more creative control and flexibility to push your designs to the next level.

Prompts are the core instructions you give the AI to tell it what kind of image to create. A prompt can be simple or detailed, but the more clearly you describe what you want—such as characters, mood, style, lighting, clothing, action, or setting—the more likely you'll get results that match your vision. Prompts should be written as natural descriptive phrases (not code or commands), and you can add modifiers like "cinematic lighting," "anime style," or "ultra detailed" to guide the output. You can also use emphasis, like putting important words in parentheses or repeating them, to boost their influence. The prompt works hand-in-hand with the model you choose—since each model has its own strengths, styles, and limitations, your wording may work better or differently depending on that pairing. Great prompts are clear, visual, and aligned with the model’s strengths.

Negative prompts tell the AI what to avoid in your image. While your main prompt guides what you want to see, the negative prompt is just as powerful for controlling what you don’t want—like blurry faces, extra limbs, bad anatomy, unwanted styles, or weird colors. By listing things like “low quality, poorly drawn hands, grainy, extra fingers, text,” you help the model steer clear of common flaws and generate cleaner, more accurate results. It’s especially useful when fine-tuning results with complex characters, realistic features, or stylized effects. Think of it as cleaning the canvas before the art even starts—negative prompts remove distractions so your main idea shines through.

Trigger words are specific terms baked into models or LoRAs that "unlock" features, styles, or characters the model was trained on. They work as key ingredients in your prompt—if a LoRA was trained to make cyberpunk robots, the trigger might be “cyberpunk unit” or a unique name like “Unit-X.” Including that word in your prompt connects directly to that training, activating the visuals or style you want. Without trigger words, those features won’t show up clearly, so they’re essential when using custom resources.

Aspect Ratio controls the shape and dimensions of your final image—like whether it’s square, portrait, or landscape. Choosing the right ratio helps fit your subject better (tall characters in portrait, wide scenes in landscape) and keeps compositions clean. It's a simple but powerful tool for making your image look more intentional.

Video Length lets you choose how long your AI-generated video will be. The default is 5 seconds, but you can extend it to 10, 20, or 30 seconds. Each increase in length requires more processing power and resources, so it costs more Aura to generate longer videos. Choosing the right length depends on how detailed or dynamic you want your scene to be—shorter lengths are great for quick loops, while longer ones allow for more complex movement or storytelling.

The NSFW toggle lets you control whether your generation allows Not Safe For Work (NSFW) content, such as nudity or adult themes. When it's off, the generation will follow safer, more general guidelines. When it's on, and if the selected model supports it, the generation can include explicit elements. Not all models are designed to handle NSFW content—some were trained strictly on safe data and will ignore or badly interpret NSFW prompts. Others are specifically trained to generate high-quality NSFW results. The toggle only works when paired with a model that supports it, so always check the model’s tags or description if NSFW output is important to your creation.

CFG Scale (Classifier-Free Guidance Scale) controls how closely the AI sticks to your prompt. A lower value (closer to 1.0) gives the AI more freedom to be creative and interpret loosely, while a higher value (closer to 10) forces it to follow your instructions more strictly. Most users find a good balance around 5–7, but you can experiment for different results. There are also three easy presets to choose from: Creative gives wild, unexpected results, Balanced offers a mix of prompt-following and creativity, and Precise sticks very closely to your prompt. These presets make it simple to switch the generation style without needing to adjust the number manually.

The sampler controls how the AI turns random noise into your final image. Think of it like the technique the AI uses to "paint" the picture step by step. Some samplers are faster (like Euler a or DDIM)—they get you results quicker but might miss fine details. Others are slower but give sharper or more detailed images (like DPM++ 2M Karras). It’s all about trading speed for quality depending on what you want. We offer two easy presets: Fast, for quick generations, and Popular, for the most commonly used high-quality settings. If you're not sure what to pick, start with Popular—it’s a solid balance for most results.

Steps control how many times the AI refines the image before finishing. More steps = more chances to improve the image, but it takes longer. Fewer steps = faster results, but they might look rough or less detailed. We give you three presets to keep it simple: Fast (around 10–15 steps), Balanced (20–30 steps), and High (up to 50 steps). If you're new, Balanced is a great place to start—it gives good quality without waiting too long.

The seed is like the "starting point" or blueprint the AI uses to generate your image. If you use the same seed with the same settings, you'll get the same image every time. That’s useful if you want to recreate or slightly tweak a result. You can choose Random to let the AI pick a new starting point every time (good for variety), or enter a Custom Seed if you want to repeat or remix a specific image. Think of random as “surprise me” and custom as “do that again.”

Priority jobs let you choose how quickly your image or video is generated. By default, every job runs at Standard speed and costs no extra Aura. But if you're in a hurry, you can pay a little extra to speed things up. Fast priority uses +10 Aura and makes your job finish about twice as fast. Blitz is the fastest option—about 3x the speed of Standard—and costs +15 Aura. It’s perfect when you want instant results or you're working on something time-sensitive. Just pick the priority that fits your needs and timeline.

You can claim 25 free Aura every 24 hours just by clicking the Claim Aura button—it’s that simple. This daily bonus is our way of saying thank you for being an active part of our community. It helps you keep creating without needing to spend, and it resets every day, so don’t forget to grab it!

The Generate button is what you click to start creating your image or video. On the right side of the button, you'll see the Aura cost, which tells you how much it will take to generate based on your settings. There’s also a reset button next to it—clicking this will wipe all your prompts and settings so you can start completely fresh. If you're curious about what’s using up your Aura, just click on the Aura counter, and a small cost breakdown will appear showing exactly what parts of your setup (like image size, steps, sampler, priority, and extras) are increasing the total cost. It’s a simple way to keep track of how your creative choices affect what you spend.