> ## Documentation Index
> Fetch the complete documentation index at: https://docs.alpacaml.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Concept Art - Character Design

> Leveraging Alpaca AI rendering for character design iteration and exploration.

<video autoPlay muted loop playsInline className="w-full aspect-video" src="https://alpaca-content-public.s3.amazonaws.com/process.mp4" />

Alpaca is a powerful tool for concept artists due to its ability to quickly and extensively explore various ideas.
During a typical concept art process, many promising ideas are left unexplored due to time constraints.
Alpaca, however, enables rapid exploration of numerous ideas, allowing for greater risk-taking while still ensuring a high-quality concept is delivered on time.

There are many different ways to integrate Alpaca in your process, our tools are made so they can be leveraged at any point in your process, be it for a simple touchup or for a full render.
In this case, and for the sake of learning, we will use Alpaca all the way from silhouetting to final render.

First, let's find a silhouette that we like.
We can sketch down a few very simple silhouettes and use Alpaca to render them in more details.
By setting Freedom to a medium value, we can draw a very simple silhouettes and let Alpaca be a bit creative to refine it.

<Frame caption="Tightening our silhouettes">
  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/silhouette.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=769524852e6695a5ebea03b749c26805" alt="AI rendering of a silhouette for character design" width="2624" height="1234" data-path="images/guides/concept-character/silhouette.webp" />
</Frame>

Now that we have a few silhouettes we like, we can start to render them to get a few details to show up.
We can render them in a single pass, or in multiple steps for more control

<Frame caption="Rendering silhouettes">
  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/silhouette-iter.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=8bd8c087449fd5883beca948aecef239" alt="AI rendering of a silhouette for character design" width="4038" height="2439" data-path="images/guides/concept-character/silhouette-iter.webp" />
</Frame>

The knight concept appears promising, so we will continue to explore this one.
What would it look like if the knight had a sword and a metal helmet?

Let's add those elements and have Alpaca render them.

<Tip>
  Here, I used Photoshop to paint the sword and helmet, I could have used
  Alpaca, but it doesn’t yet have soft brush which I wanted to use here.
  Thankfully going back and forth between Photoshop and Alpaca is very easy, as
  we can simply copy/paste the canvas from one to the other with `ctrl+c/ctrl+v`
  on Windows or `cmd+c/cmd+v` on mac.
</Tip>

<Frame caption="Adding in details to our character">
  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/gen-mask.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=dd529ab6df49f2c5fe5803415c5836e2" alt="Drawing in details to our character" width="2175" height="960" data-path="images/guides/concept-character/gen-mask.webp" />
</Frame>

Since we only want to render the helmet and sword without changing the rest of the image, we can use the **Generation Mask** tool.
This tool allows us to highlight the area of the image we want to render, leaving the rest unmodified.

To use the **Generation Mask** tool, select it from the toolbar or simply press `L`. You can then highlight the area to render using the brush.

{" "}

<img width="256" src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/gen-mask-btn.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=748a65c6f3b3bfd2eca87dd5eda0a19d" alt="Generation Mask button" data-path="images/guides/concept-character/gen-mask-btn.webp" />

<Frame caption="Rendering a specific area of our image with Generation Mask">
  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/mask.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=0fd09880acfb1031ffabb9aa90c03c22" alt="Drawing in details to our character" width="957" height="960" data-path="images/guides/concept-character/mask.webp" />

  {" "}

  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/orig.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=4794e1ee8c2894153018a9598f9b5633" alt="Drawing in details to our character" width="960" height="960" data-path="images/guides/concept-character/orig.webp" />
</Frame>

So far, we have explored in a pretty straight line, refining a single idea more and more.
Let’s see if we can spice things up a little bit by rendering some variations of our concept with different twists.

To do that we can set **Freedom** to a fairly high value, around 30 or 50. This tells Alpaca to not be overly concerned about respecting our current input precisely, leaving space for interesting things to emerge. We can set **Change** around 75 or so.

Now, we can reuse the same prompt we had earlier (`knight with a sword, fantasy character, blank background`) , but add a specific modifier we are interested in exploring in front of it.

<Frame caption="Generating variations of our concept">
  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/variations.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=4964bbe62366f83c8d5de4633c549623" alt="Drawing in details to our character" width="2944" height="1167" data-path="images/guides/concept-character/variations.webp" />
</Frame>

This is nice, the dragon armor variation is pretty cool. But can we push this idea further?
Alpaca has another functionallity that we haven’t yet leverage in this project: `Style References` .

With **Style References** we can use other images to guide the style of our renders.

Let’s see if we can use it to make our render a bit more fiery.

We are going to use the two following reference images (on **Free** and **Standard** plan you can only use a single reference image at a time, but that will still work just fine!)

<Frame caption="Style References">
  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/fire.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=796b14c3ee03383403e630854e0e1770" alt="Style reference for ai rendering 1" width="1106" height="622" data-path="images/guides/concept-character/fire.webp" />

  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/fire2.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=4ba947044a4e95b480156caa0dc8617b" alt="Style reference for ai rendering 2" width="1016" height="610" data-path="images/guides/concept-character/fire2.webp" />
</Frame>

We will upload the images to the style references tool and set the reference strength to approximately \~20. A higher reference strength might reproduce some details from the reference image in the output, which we want to avoid in this case.

We will maintain a high value for **Change** (between 80 and 100) and afford some **Freedom** to Alpaca, allowing it to render new details like flames (around 20 or 30).

<Frame caption="Using Style References in Alpaca">
  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/ref.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=a962893e4314c299aae282cc616f73a0" alt="Using Style References in Alpaca" width="1002" height="974" data-path="images/guides/concept-character/ref.webp" />
</Frame>

Here are a few examples of renders we get:

<Frame caption="Using Style References in Alpaca">
  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/ref-variations.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=aa34a22a4c04f8108178f1785b91c295" alt="Using Style References in Alpaca" width="3104" height="1016" data-path="images/guides/concept-character/ref-variations.webp" />
</Frame>

This is pretty nice. But we're going to stick with one of our earlier prototypical knights and make a few more modifications to reach our final design. These will include giving our knight glowing eyes, giving them a larger broadsword, and adjusting the armor to be a bit spikier.

These changes can be easily made using the **Generation Mask** technique we've previously discussed.

<Frame caption="Final iteration on our concept">
  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/pre-final.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=dcbf8537f97d65030d5bee0f578c8596" alt="Using generation mask in Alpaca" width="960" height="960" data-path="images/guides/concept-character/pre-final.webp" />

  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/pre-final-iter.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=e8cc8f4c2adcb25d4a8935f95c938464" alt="Using generation mask in Alpaca" width="960" height="960" data-path="images/guides/concept-character/pre-final-iter.webp" />
</Frame>

And here is our final render

<Frame caption="Final render">
  <img src="https://mintcdn.com/alpacaml/vUSqxKC-ubb-0IQ6/images/guides/concept-character/final.webp?fit=max&auto=format&n=vUSqxKC-ubb-0IQ6&q=85&s=a3dd46820119448d90f7daef45ce5e0f" alt="Final rendering" width="1645" height="1381" data-path="images/guides/concept-character/final.webp" />
</Frame>
