← Maki

How to Make a Mood Board: Step-by-Step Guide for Artists

Maki blog · 2026-10-28

A mood board is a collection of images, colors and notes that shows what a project should feel like before you start drawing. You make one by writing a short brief, gathering more images than you need, cutting the pile down to the strongest 20 to 40, arranging them, and annotating what matters. This guide walks through each step with examples from illustration, concept art and character design.

In short: Write a one-line brief with three to five keywords. Collect 50 to 80 candidates from your folders, Pinterest and your own photos. Cut to 20 to 40 images, arrange them in a grid or hierarchy, pull a color palette, annotate why each image earned its place, and name the board so you can find it in six months. The whole process takes one to two hours for a personal project.

What a mood board is (and what it is not)

A mood board is a single view of the visual direction of a project. It answers one question: what should this look and feel like? It is not a plan for composition, anatomy or rendering. It is not a to-do list. It sits before thumbnailing, and it should be finished before thumbnails start.

Think of it as a contract with yourself, or with a client. When you are three days into a painting and everything looks wrong, the board tells you what "right" was supposed to be. That is its real job. Not inspiration. Correction.

What goes on a mood board

Anything visual that carries information about direction:

What does not belong: finished work of your own from the same project, client feedback documents, anything you would feel awkward explaining to the person who made it. If an image is only there because it is pretty, and it does not serve the brief, it will confuse the board.

Mood board vs reference board vs vision board

These three get used interchangeably, and that causes problems, because each one answers a different question at a different stage.

A mood board is early and directional. It defines feeling, color and shape language for one specific project. It is usually 20 to 40 images and it is finished when the direction is agreed.

A reference board is late and factual. It holds the specific things you need to draw accurately while working: the exact armor joint, the way leather creases at the elbow, the facade of that particular building. A reference board grows during the project and it is never really "done". Many artists keep a general reference library that is not tied to any project at all, and pull from it when making mood boards.

A vision board is personal and long-term. It is about goals: the kind of work you want to be making in two years, the clients you want, the style you are building toward. It has no deadline and no client.

The confusion costs real time. If you mix project reference into a mood board, the board stops communicating direction and becomes a pile of stuff. Keep them separate. Make the mood board first, agree on it, then build the reference board underneath it.

Takeaway: A mood board defines direction for one project. Keep it separate from working reference and from long-term goals, or it stops doing its job.

Before you collect: write a one-line brief

Most bad mood boards were doomed before the first image was saved. The failure was not taste. It was the absence of a sentence.

Before you open anything, write one line that states what the board is for. For example: "A lonely lighthouse keeper's home, warm interior against a hostile grey sea, painted children's book style." That single line will delete half of your candidate images for you, before you even save them.

Pick three to five keywords

From that line, extract three to five keywords. For the lighthouse example: warm interior, grey sea, worn wood, quiet. Write them at the top of the board or in a corner. These words are your filter for the next hour of collecting.

Three to five is the working range. Fewer than three and the direction is too vague to filter anything. More than five and you are making two boards. A board trying to be "cozy AND epic AND minimalist AND ornate" will please no one, including you.

Keywords also make the board self-explaining. When you share it, the viewer reads four words and immediately understands the organizing idea. You will not have to narrate.

Decide who the board is for

Boards change shape depending on the audience, and deciding this early saves a second pass later.

Be honest about which one you are making. The most common waste is polishing a personal board to client standard, or sharing a personal mess with a client and spending the call explaining it.

Takeaway: Write the one-line brief and three to five keywords before collecting anything. Decide who reads the board, because that determines how clean it needs to be.

Step 1: Gather images from where you already look

Now you collect. The goal at this stage is volume with judgment: enough candidates that cutting later is easy. Expect to gather 50 to 80 images for a board that will end at 20 to 40. Do not curate yet. Curating while collecting is the slowest way to do this.

Set a time limit. Forty-five minutes to an hour is enough for a personal board. If you are still gathering after two hours, you are procrastinating, and the pile will only get harder to cut.

Your own folders and screenshots

Start with what you already own. Most artists have years of saved reference sitting in folders called "ref", "inspo" and "stuff". This material is better than a fresh Pinterest session for two reasons: you already chose it once, and you already know where the good parts are.

Pull your existing folders into your board tool. If you use a desktop app, you can usually drag whole folders in from Explorer or Finder rather than importing file by file.

Screenshots count. That lighting study you screenshotted from a film last month is a candidate. Your camera roll of textures from a walk is a candidate. Old work of your own, from before this project, can define what you want to avoid; keep those in a separate corner if so.

Pinterest and the wider web

Pinterest is where most artists start, and it is good for one thing: volume and surprise. Search your keywords, then follow the related searches Pinterest suggests. Save generously. You are going to delete most of them.

The weakness of Pinterest is that everything there is already curated by popularity, so boards drift toward the same 200 images everyone else has. Break out of it deliberately:

If you have a Pinterest board already, you do not need to re-save anything by hand. Maki's browser extension imports a whole Pinterest board in one click, up to 1,000 pins, and turns it into its own Maki board that you can then cut down. That turns an hour of re-saving into about a minute.

Photos you take yourself

Your own photos are the strongest material on any board, because nobody else has them. A phone photo of peeling paint on your street beats the 40th Pinterest image of peeling paint. It is specific. Specific is what makes a board feel like a direction instead of a trend.

You do not need a good camera. You need ten minutes and your phone. For the lighthouse board: photograph wood grain on old doors, grey skies from your window, any warm lamp in a dark room. Shoot details, not scenes. Details transfer into paintings; scenes just get copied.

Photograph your own sketches too. If you have thumbnail ideas already, a photo of the three best ones belongs on the board, marked as yours. It anchors the direction in your own hand.

Takeaway: Gather 50 to 80 candidates in under an hour, starting with your own folders and photos, then Pinterest and archives for volume. Do not curate yet.

Step 2: Cut the pile down

This is the step most artists skip, and it is the difference between a board that communicates and a board that wallpapers. Cutting is not loss. Every image you delete makes the remaining ones louder.

Work fast. Give yourself 15 to 20 minutes for a pile of 60. If you linger on individual images, you are negotiating with yourself, and you will lose.

Why 20 to 40 images is enough

Here is the concrete answer to a question almost nobody addresses: how many images should a mood board hold?

For a working artist's board, 20 to 40 images. That range is not arbitrary.

There are exceptions. A board for a single character's face might live at 15. A world-building board for an entire game region might justify 60. But if you cannot say why your board is an exception, cut to the range.

A practical test: show the board to someone for ten seconds, then take it away and ask what the project feels like. If they cannot answer, the board is too big or too scattered. 20 to 40 images passes this test reliably; 80 does not.

What to delete first

Delete in this order. It is faster than deciding image by image.

  1. Exact duplicates and near-duplicates. If two images say the same thing, keep the better one. One exception: deliberate repetition of a core idea, like those three grey seas, can stay if each adds something.
  2. Off-brief images. Anything that does not serve your keywords. The beautiful image that has nothing to do with the lighthouse goes, no matter how good it is. Save it somewhere else if it hurts.
  3. Images that only demonstrate skill. A rendering masterclass with no mood content communicates "I wish I could paint like this", which is not a direction.
  4. Trend filler. The images you have seen on fifty other boards. If it arrived from a generic Pinterest search and says nothing specific, cut it.
  5. Anything you cannot explain. If you cannot say in one sentence why an image is on the board, it is decoration. Cut it or write the sentence.

Expect to delete roughly half of what you gathered. That is normal. If you deleted almost nothing, you were too attached while collecting, and the board will be weaker for it.

Takeaway: Cut to 20 to 40 images. Delete duplicates, off-brief images and skill demos first. Half your pile should go.

Step 3: Arrange the board

Arrangement is what turns a pile into a board. The same 30 images arranged differently tell different stories. You are controlling where the eye lands first, second and third.

There are three working layouts. Pick one deliberately rather than drifting into whatever the drag-and-drop produced.

Grid layout

Every image the same size, in rows and columns. Nothing overlaps.

Use it when the board needs to look organized and neutral: client presentations, team boards, anything that will be viewed without you. A grid says "these images are equal contributors to a direction". It is also the fastest layout to build, since most tools will snap and distribute for you.

The risk of a grid is monotony. Thirty equal rectangles can flatten a strong direction. Fix it with content, not layout: make sure the grid contains variety in subject and color, so the rhythm comes from the images themselves.

Hierarchy layout

One or two large images anchor the board, with smaller supporting images around them.

Use it when there is a hero idea. The large image is the thesis: the keyframe feeling, the main character's silhouette, the defining environment. Everything smaller explains or supports it. This is the layout to choose when the board must persuade, because the viewer gets the point before reading anything.

A hierarchy board for the lighthouse project might have one large image of a warm window seen from outside in the rain, with clusters of interior details, sea studies and material swatches around it. One look and the story is there.

Clusters by color, shape or idea

Images grouped by theme, with space between groups. Groups can overlap at the edges, which is fine and often useful.

Use it for your own working boards and for world-building, where several sub-directions coexist. A game region board might cluster into "architecture", "foliage", "lighting", "props" and "color". You can then delete or grow a whole cluster without disturbing the rest.

Clustering is also how you discover problems. If one cluster has 25 images and another has 3, your research is lopsided, and you know exactly where to spend the next 15 minutes.

Whichever layout you choose, leave negative space. Boards packed edge to edge read as clutter regardless of content. Space between groups is what makes groups legible.

Takeaway: Choose your layout on purpose: grid for neutral presentation, hierarchy to persuade, clusters for working boards with sub-directions. Leave space between groups.

Step 4: Pull a color palette

A mood board without a palette is half a board. Color is usually the first thing a viewer registers, often before subject matter, and it is the thing most likely to drift during a long project. Pulling the palette now, from the board itself, locks it.

Sampling colors from your images

Go through the board and sample the colors that repeat. Not the colors you like. The colors that keep showing up, because repetition across your chosen images is the board telling you what the palette actually is.

Collect somewhere between 5 and 8 swatches. Fewer than 5 and you have a duotone, more than 8 and you have a sampler, not a palette. A useful structure:

Place the swatches as a strip on the board itself, in a corner or along an edge. Do not keep the palette in a separate file. A palette that lives on the board travels with the board, gets seen in every review, and survives being shared as a single image.

Most tools have an eyedropper. Your operating system probably has one built in too. Sample from your own photos where possible; they tend to give less processed, more usable colors than heavily graded film stills.

Checking the palette in grayscale

Before you commit, convert the board to grayscale and look again. This takes a minute and catches problems that color hides.

What you are checking is value structure. If your dominant colors all convert to the same middle gray, your palette has no value range, and paintings made from it will feel flat no matter how good the hue choices are. You want a spread: dark masses, mid-tones and lights, each doing distinct work.

The classic failure: a palette of five beautiful muted blues that all sit at the same value. In color it looks sophisticated. In grayscale it is one flat tone. Every painting built on it will struggle with depth.

Fix it by swapping one or two swatches for lighter or darker neighbors of the same hue. The mood survives; the structure appears.

Takeaway: Sample 5 to 8 swatches from the colors that repeat on your board, put the strip on the board itself, and check the whole thing in grayscale before committing.

Step 5: Annotate the board

Annotation is what separates a board you can use from a board you can only look at. In three months you will not remember why that specific door hinge photo is on the lighthouse board. Your future self needs notes.

Notes that explain why an image is there

Every image that is not self-explanatory gets a short note. Not a description of what the image shows. The reason it earned its place.

Bad note: "Old door." Good note: "Door color, not door shape. Slightly warmer than this." Bad note: "Sea reference." Good note: "This amount of foam. Less looks calm, more looks storm."

Notes are also where you record what to ignore. "Like the sky, ignore the boat." This is the single most useful type of note on a shared board, because it stops teammates extracting the wrong thing from an image.

Keep notes short. A sentence. If a note needs a paragraph, the image is doing too many jobs and one of you should be cut.

Tags that help you find it later

Notes are for humans reading the board. Tags are for you searching your library later.

Give images a small set of consistent tags: project name, subject, and one or two attributes. "Lighthouse, wood, interior." The discipline that pays off is consistency: decide your vocabulary once and stick to it. If some images are tagged "interior" and others "indoors", your future searches find half of what you own.

Tags and notes also make boards searchable in ways file names cannot. If your tool reads notes when you search, you can find that door photo by typing "warmer door" years later.

Takeaway: Note why each image is there and what to ignore, in one sentence. Tag with a consistent vocabulary so the board stays findable years later.

Step 6: Share it and get feedback

A mood board that is never shared still has value, but a shared one gets better faster. Feedback at the board stage is cheap. Feedback at the finished painting stage is expensive. Move the conversation earlier.

Exporting a PNG or PDF

For most purposes, export the whole board as a single PNG. One file, one glance, nothing to install for the viewer. PNG keeps text and flat color crisp, which matters when your board has notes and swatch strips.

Use PDF when the board needs to travel into a document pipeline: client decks, production bibles, print. PDF also lets you attach source links as proper annotations in some tools, which matters for the licensing section below.

Name the export the same way you name the board, with the date: "lighthouse-keeper-2025-03.png". Six months from now, a folder of exports called "board-final-final-2.png" is a small disaster.

Walking someone through the board

When you present the board live, do not narrate every image. Spend your minutes on structure:

  1. Read the one-line brief and the keywords aloud. Ten seconds.
  2. Point at the hero image or the dominant cluster. Say what it is doing.
  3. Walk two or three supporting groups, one sentence each.
  4. Show the palette strip and name the dominant colors.
  5. Stop talking. Ask what feeling they get, and listen for words that differ from your keywords. Those gaps are the real feedback.

If the viewer's summary of the board matches your brief, the board works and you can start thumbnails. If it does not, the board needs fixing, and you just found that out before spending a week painting.

One more habit worth building: keep the board visible while you work. Many desktop reference tools can keep a window on top of your painting app, so the direction is always in view rather than buried behind your canvas. It sounds minor, but direction drifts when the board is hidden.

Takeaway: Export one PNG named with the project and date, present structure rather than image-by-image narration, and keep the board on screen while you paint.

Mood board examples by discipline

The steps above are universal, but each discipline stresses different parts of the process. Here is how the weight shifts.

Illustration

For illustration, the board is mostly about style, palette and lighting, because subject matter usually comes from the brief itself. A children's book board might hold 25 images: texture treatments from other illustrators, photos of the actual materials the scene involves, lighting references for the key emotional beat, and a palette strip.

One trap specific to illustration: boards drift toward the work of a handful of famous illustrators, and then your finished piece reads as an imitation of one of them. Deliberately mix in photography, traditional media scans and non-illustration fine art. Your style should be a synthesis the board suggests, not a single artist it points at.

Concept art and game art

In concept art the mood board is often called something else, like a kickoff board or a visual target, but it does the same job, and it usually comes with a callout sheet of specific references for architecture, materials and technology.

Game art boards are bigger and more structured. A region board might legitimately hold 60 images across clusters: environment, architecture, props, foliage, lighting moods, and a "do not do" cluster of the direction the team rejected. That rejection cluster is genuinely useful. It prevents relitigating the same argument every sprint.

For game work, keep the board alive during production rather than archiving it after approval. New reference will arrive constantly. If your tool links to live project folders, the board updates as the team drops files in, and it stays current without anyone maintaining it.

Character design

Character boards are the most personal and the most prone to bloat, because "inspiration for a character" can include almost anything. Discipline comes from structure. A working character board has distinct zones:

For a character board, cut harder than usual. Fifteen to 25 images is often enough, because a character direction is narrower than an environment direction. And annotate aggressively: "the confidence, not the armor" is the note that keeps a design on track through twelve revisions.

Brand and logo work

Designers make mood boards too, and if you take on logo or brand work, the process is the same with different content. A brand board leans on typography specimens, historical design eras, material and print finishes, and competitor screenshots, alongside the usual palette.

The difference is the audience. Brand boards are almost always shown to non-designers, so the hierarchy layout wins: one strong hero image that captures the brand's temperament, supported by tight clusters. And the "do not do" cluster matters even more here, because clients will happily pull the board toward their competitor's look if the boundary is not visible.

Takeaway: Shift the emphasis by discipline: style and palette for illustration, structure and live updates for game art, hard cutting and zoning for characters, hierarchy and boundaries for brand work.

Digital vs physical mood boards

Both are legitimate. They do different things well, and many artists use both: physical for early exploration, digital for the board that actually gets used.

Physical boards

A corkboard, a wall, or a stack of printed images pushed around a table. Physical boards have real advantages: printed images have a material presence that screens flatten, you can hold two prints side by side at arm's length, and rearranging means physically moving things, which makes composition decisions slower and more deliberate.

The costs are just as real. You cannot search a wall. You cannot back it up. Sharing it means photographing it, which loses the scale that made it good. And space limits the library: a physical board of 40 images is a commitment of wall, a digital one is a Tuesday.

Use physical for the first hour of a project, when you want to feel the direction, then photograph the arrangement and rebuild it digitally as the working board.

Online whiteboards

Browser-based boards like Milanote and Pinterest itself are the most common digital option. Their strength is access: nothing to install, works from any machine, easy to share a link, and real-time collaboration if a team is scattered.

The costs: they need an account, they need a network connection, and your boards live on someone else's servers under someone else's terms. Subscription pricing adds up over years for what is, functionally, a pile of images. And if the service changes direction or shuts down, your library goes with it. For artists whose reference collection is effectively a career asset, that is a real risk, and it is worth thinking about before committing years of collecting to any single web service.

Desktop apps that work offline

The third option is a desktop app that stores boards as files on your own disk. No account, no subscription, works on a plane or in a studio with bad wifi, and your library is yours in the plainest sense: it is in a folder, and you can back it up like any other folder.

This category includes tools like PureRef and Eagle, and Maki sits here too. The trade-offs run opposite to web tools: you give up browser access from any device and real-time collaboration, and you gain speed and ownership. A board of a few hundred images opening instantly, offline, on your own machine, changes how willing you are to actually use reference day to day. If you are comparing options, we have written guides on PureRef alternatives, Eagle alternatives and Milanote alternatives elsewhere on this site.

Takeaway: Physical boards are best for early exploration, web boards for distributed teams, and offline desktop apps for artists who want speed, ownership and no dependency on a service staying alive.

Common mood board mistakes

Almost every bad mood board fails in one of three ways. All three are cheap to fix if you catch them.

Starting without a brief

Collecting before deciding what the board is for produces a pile with no filter. The pile grows, nothing gets cut because there is no criterion to cut against, and after three hours you have 200 images and no direction. The fix is the one-line brief from earlier in this guide. If you are already deep in a pile with no brief, stop, write the line, and then cut. It is never too late to write the sentence.

Collecting everything

The second failure is treating the board as a library. Libraries are valuable, but they are a different artifact. A mood board is an argument, and arguments need editing. The symptom is a board past 60 images where nothing stands out and you cannot find the palette. The fix is the 20 to 40 range and the deletion order from Step 2. If the material is genuinely good, move it to a reference library instead of the board, and keep the two separate.

Losing where an image came from

This one has consequences beyond a weak board. If you cannot say where an image came from, you cannot check its license, you cannot credit it if the work gets published, and you cannot find it again when you need a higher resolution version.

The fix is a habit, not a tool: when you save an image, save its source with it. Paste the URL into the image's note field. If your tool tracks the source page when you clip from a browser, use that. The Save to Maki extension, for example, records where an image came from when it sends it to a board, so the source travels with the file.

Be aware of what you are doing with other artists' work, too. A mood board is usually private or internal, which is a normal, fair use of reference. The line gets crossed when a board image gets traced closely, or when the board itself is published commercially. If a board will be printed in an art book or shown in a paid tutorial, you need permission for the images that are not yours. Track sources from day one so that this is a lookup rather than an archaeology project.

Takeaway: The three failures are no brief, no editing and no source tracking. Write the line, cut to the range, save the URL with the image.

Keeping old mood boards useful

This is the part nobody writes about, and it is where months of work quietly disappears. You will make dozens of boards this year. In a year, you will need one of them again, and the difference between a five-second find and an hour of folder archaeology is decided by habits you set up today.

The scenario is always the same. A client from two years ago emails: "Can we do the sequel in the same feel as the first one?" Or you start a personal project and remember you already researched exactly this in 2023. The research exists. The question is whether you can reach it.

Naming and dating boards

Give every board a name a stranger could file correctly. The convention does not matter; having one does. A workable pattern is project-subject-date: "lighthouse-interior-2025-03", "clientX-characters-2024-11".

Two rules make names work. First, the project name comes first, because you search by project more often than by subject. Second, the date goes on the end in year-month format, so boards sort chronologically in any file listing and you can tell at a glance which of the three "forest-exploration" boards is the newest.

Resist clever names. "The sad robot one" is funny on the day and useless in a year. If you must, put the joke in the board's notes, where search can still find it, and keep the file name boring.

Searching by the words inside images

Names and tags fail in a specific way: you cannot tag for what you did not know would matter. You tagged the lighthouse board "interior, wood, sea". What you remember two years later is that it had a photo of a specific enamel kettle. No naming convention saves you.

This is where search that reads inside images changes the game. If your tool can recognize the text inside images, you can find that photo of a shipping label, a shop sign or a book page by the words it contains. Maki does this offline: search reads file names, notes, table cells and the words inside the images themselves, so a board surfaces by its actual content. When you are comparing tools, this capability is worth checking for specifically, because it is rare and it is the difference between a library you search and a library you browse.

It also changes what you save. Screenshots of tweets, forum posts, tutorial steps and annotated diagrams become first-class searchable reference instead of dead images. A library where the words inside images are findable is a genuinely different asset from a folder of pictures.

Backing up your library

Your reference library is a career asset. Years of collecting, curating and annotating, gone in one disk failure. Back it up like your portfolio.

The minimum is the 3-2-1 pattern: three copies of the data, on two different types of storage, with one copy off-site. In practice: the working library on your machine, a copy on an external drive, and a copy in cloud storage or at another location. Automate it, because a backup you have to remember is a backup you will eventually forget.

Check whether your tool helps or hinders this. If boards live as plain files on your disk, any backup software handles them, and you can verify a backup by opening the folder. Some tools also keep their own daily backups, which covers the "I deleted a board yesterday and regret it today" case that ordinary backups miss.

Test your restore once. A backup that has never been restored from is a hypothesis, not a backup. Pick a slow afternoon, restore one board to a different folder, and confirm it opens. Ten minutes now, or an honest conversation with a client later.

Takeaway: Name boards project-first with a year-month date, use a tool that searches the words inside images, and run 3-2-1 backups with a tested restore. Old boards are only useful if you can find them.

Key takeaways

FAQ

How many images should a mood board have?

Between 20 and 40 for a working artist's board. Under 20 usually means the direction is under-researched; over 40 and nothing stands out. A single character board can go as low as 15, and a world-building board can justify 60, but those are exceptions you should be able to defend.

What is the difference between a mood board and a reference board?

A mood board is early and directional: it defines feeling, color and shape language for one project, and it is finished when the direction is agreed. A reference board is late and factual: it holds the specific things you need to draw accurately, and it grows throughout the project. Keep them separate.

Can I make a mood board for free?

Yes. A folder of images and a free tool will do it. Pinterest boards are free, and several desktop tools are free or have free tiers. If you outgrow free tools, paid options exist at every price point, including one-time purchases like Maki at $19 for up to two computers with a 30-day refund, at the time of writing.

Should my mood board be digital or physical?

Both work. Physical boards are good for the first hour of exploration, when you want to feel the direction and move prints by hand. Digital boards are what you will actually use during production, because they are searchable, shareable and back-up-able. Many artists do both.

How long should making a mood board take?

One to two hours for a personal board: about an hour to gather, 15 to 20 minutes to cut, and 20 to 30 minutes to arrange, pull a palette and annotate. If you are past three hours, you are procrastinating, and the pile is getting harder to cut, not easier.

Do I need to credit the images on my mood board?

For private or internal use, no. Mood boards are a normal, fair use of reference. If the board itself will be published, printed in an art book, or used in paid material, you need permission for images that are not yours, which is why recording the source of every image from day one matters.

Try Maki free for 30 days

Windows 10 and 11, macOS 11 and later. No account, no card. Then $19 once, with a 30-day refund.

Related

Reference manager for artists Moodboard app for Windows PureRef vs Maki More from the blog