Is there a Claude Code skill for AI mood boards?
Yes, Wireflow's AI mood board generator skill for Claude Code turns a brief into four matching images plus a palette and tone note, for 43 credits per run.
Yes, Wireflow's ai-mood-board-generator skill makes an AI mood board in Claude Code from a campaign brief, a few mood words and written reference notes: four matching images plus a direction note with the palette and tone. Claude drafts the inputs, shows the credit cost (43 credits per run) and waits for your yes, then one run of the skill's Wireflow app writes the note and draws a hero picture, a texture close-up, a place and a detail in one look.
The skill is meant for an agency, freelancer or brand team that needs to show a client where a campaign could go visually, before anyone writes scenes or books a shoot. Your brief goes to Wireflow through the Wireflow MCP connector, signed in as you. Without the connector, the skill still writes the direction note and four image prompts for you, ready for a designer or any image tool.
For a shot-by-shot sequence, use ai-storyboard-generator instead, and to apply a brand that already exists, use brand-consistent-images. You can read the instructions Claude follows in the mood board skill's GitHub folder.
How do I install the mood board skill in Claude Code?
Install the Wireflow plugin, which gives you every Wireflow skill plus the Wireflow connector:
/plugin marketplace add wireflowINC/skills
/plugin install wireflow-skills@wireflow
Claude Code shows the plugin's details first, so you can review it and choose its scope. Then run /mcp, choose wireflow and sign in with your Wireflow account. The skill shows up as /wireflow-skills:ai-mood-board-generator, and Claude also picks it up from plain requests.
For this skill alone, use the skills CLI, or copy the skills/ai-mood-board-generator folder into ~/.claude/skills/ (all projects) or .claude/skills/ (one project). Then connect Wireflow yourself and run /mcp to sign in. The connector uses OAuth, so there is no API key to copy.
npx skills add wireflowINC/skills --skill ai-mood-board-generator
claude mcp add --transport http wireflow https://www.wireflow.ai/api/mcp
To check that it worked, ask Claude for a mood board. With Wireflow connected, the cost quote shows your credit balance. If the Wireflow tools are missing, the skill says so in one sentence, gives you the setup steps and writes the direction note and image prompts by hand instead.
What do I tell Claude to make a mood board?
Tell Claude about the client, the campaign and the feeling you are after, in plain words, and say how many routes (visual directions) you want. For example:
Make a mood board for a client pitch. The client is a small sea salt brand
launching a summer gift range for food lovers. It should feel salty, warm and
handmade, never glossy or corporate. Lean on 1970s seaside kiosks, sun-faded
paper and rough ceramics. Give me two routes to compare.
The skill is meant for requests like this one: a visual direction, a look and feel, a style exploration, a moodboard for a client pitch, or two or three routes to compare before a campaign. A good request covers the four things the skill asks for:
- The client and the campaign: what it sells, to whom, and why now.
- How it should feel, in a few words, and what it must not feel like.
- Anything to lean on: eras, places, materials and photographers' habits, described in plain words.
- One route or several. Each route is one run.
You can also paste a client brief: Claude pulls the answers out of it and shows you what it took. Leave out unreleased client details you would not want on a third-party service, and Claude asks when it is unsure.
From there, the skill:
- Fills in three inputs for each route: the campaign brief (up to 2,000 characters), three to six mood words, and reference notes.
- Shows the routes in a table with their brief, mood words and reference notes, for you to edit.
- Quotes the credit cost with your balance, and runs nothing before your yes to that table and that cost.
- Calls
run_apponce per route and pollsget_executionevery 5 seconds until the run finishes. The skill's test took about a minute: image 1 comes first, and the other three wait for it. - Saves the files, looks at all four images and reports back.
If you edit the table, keep all three inputs filled. A missing one falls back to the board's saved demo, an oat milk launch, which would leak into your result. With no references to give, the skill sends "No references, choose freely".
What do I get from one mood board run?
You get a direction note and four landscape 4:3 JPG images with no text, saved to ./wireflow-outputs/YYYY-MM-DD/mood-boards/route-slug/ with the date and the route's slug filled in.
| File | What it is | Format |
|---|---|---|
direction-note.md |
The idea, a Palette with hex codes, the Tone and an Avoid line | Text |
1-hero.jpg |
The hero picture | JPG, 3072x2304 in the skill's test |
2-texture.jpg |
A texture close-up | JPG, 2560x1920 |
3-place.jpg |
A place | JPG, 2560x1920 |
4-detail.jpg |
A detail | JPG, 2560x1920 |
Inside the app, Claude Haiku 4.5 sets the direction by writing the note and four image prompts, and Seedream 5 Lite draws the images. Image 1 is drawn from its own prompt. Images 2 to 4 use image 1 as the reference for medium, palette and light, each with its own subject.
Before calling it done, Claude looks at all four images and names every miss, file by file:
- The four should read as one look: same medium, same light, the palette from the note.
- No stray writing. Small marks on dials, labels, signs and packaging are where it slips in.
- Nothing from the Avoid line, and nothing a client could mistake for a real brand.
- Object details can drift between images, such as a ceramic cup in one and an enamel mug in the next. That is fine for direction, and Claude names it.
The report gives a table with each route, file and link, the direction note, the credits used (your balance before minus after) and one line on how the routes differ. Treat the images as direction references, not final art. Claude then offers another route on the same brief, a next step for scenes once you pick a route, and the board in Wireflow.
Can the mood board skill use my reference images?
No, the inputs are text only and the app cannot see pictures, so describe your references in words: eras, places, materials and things to avoid. Claude turns named brands, artists and films into descriptions before they go in, for example "1960s travel posters" rather than a designer's name.
The skill's guardrails also leave out real people, celebrities, other brands' logos and the named style of a living artist.
Can I make two or three mood board directions to compare?
Yes, each direction is a route and each route is its own run, so two routes means two runs side by side. Claude makes the routes truly different by changing the world or the era, not only the colours, and quotes them together before anything runs, for example:
2 routes x 43 credits = 86 credits for 8 images and 2 notes. Balance: 1,240.
If your balance is short, Claude stops and says so. Adding another route on the same brief later gets its own quote.
Can I change the image size or add a fifth image to the mood board?
Yes, by opening the skill's board in Wireflow: the skill's app makes four landscape 4:3 images, and at the end of a run Claude offers you the board to change the image size or add a fifth image.
What happens if a mood board run fails?
Claude stops that route and quotes the error when a run ends as FAILED, CANCELLED or TIMEOUT. A json_extract_field error means the direction step wrote broken JSON, so Claude offers one retry, which runs only after a new yes. When Wireflow refuses or a tool errors, Claude quotes the exact message and offers send_feedback if that tool is available.
How many credits does one AI mood board cost?
One mood board costs 43 credits per run: 10 per Seedream 5 Lite image plus 3 for the direction step, as measured on 2 Nov 2026.
| What | Credits |
|---|---|
| The brief, the routes and the quote, planned in Claude | 0 |
| Each Seedream 5 Lite image | 10 per image |
| The direction step | 3 per run |
| One route: four images and a direction note | 43 per run |
| A redo or another route | 43 per run, only after your yes |
Generation needs a Wireflow account with credits. See pricing for what each plan includes.
Does the mood board skill work in Cursor, Codex or Claude Desktop?
Yes, the skills CLI also installs it in Cursor, Codex, Gemini CLI and other agents that read SKILL.md, but each agent must be connected to the Wireflow MCP server before the skill can make images. The skill is written for Claude Code and Claude, and Wireflow notes that it has not tested its skills in other agents itself.
In Claude.ai or Claude Desktop, add https://www.wireflow.ai/api/mcp as a custom connector. Saving files needs a shell.
Connection steps are on the Cursor and Codex pages, and for every client in Use Wireflow from your agent. To build and run your own Wireflow workflows over the REST API instead of an agent, see the REST API reference. More Wireflow skills are on the agent skills page.
Related
- Skills
Yes, Wireflow's public agent skills make images, videos, product shots and ads in Claude Code, and the skills CLI installs them in Cursor and Codex.