---
title: "Is there a Claude Code skill for AI social media graphics?"
description: "Yes, Wireflow's AI social media graphics skill for Claude Code turns a quote, stat or announcement into one post in three sizes, at 57 credits per run."
updated: 2026-11-14
---

Yes, Wireflow's `ai-social-media-graphics` skill turns a quote, a stat or an announcement into one branded text post inside Claude Code, drawn in square, portrait and story sizes. It trims the words, lets you approve every piece of text, quotes the credits and waits for your yes, then renders the post with Ideogram 4.5 and saves all three PNGs to `./wireflow-outputs/`.

The words and the layout are settled in Claude, and the render runs as the published Wireflow app `skill-ai-social-media-graphics` over the Wireflow MCP connector, signed in as you. Without the connector, the skill still writes a post sheet and a design spec a designer can build. The skill's files are in [its folder on GitHub](https://github.com/wireflowINC/skills/tree/main/skills/ai-social-media-graphics).

Use it for a quote card, a stat post, an announcement graphic, an event or launch post, or the same post sized for feed and stories for a client. For a multi-slide story, the skill points you to `ai-instagram-carousel-generator`, and for paid product ads to [ai-ad-creative-generator](/docs/skills/ai-ad-creative-generator).

## How do I install the social media graphics skill in Claude Code?

Add the Wireflow plugin, which installs every Wireflow skill together with the Wireflow connector:

```text
/plugin marketplace add wireflowINC/skills
/plugin install wireflow-skills@wireflow
```

Then run `/mcp`, choose `wireflow` and sign in with your Wireflow account. The skill appears as `/wireflow-skills:ai-social-media-graphics`, and Claude also picks it up from plain requests.

To add only this skill, use the skills CLI, or copy the `skills/ai-social-media-graphics` 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.

```bash
npx skills add wireflowINC/skills --skill ai-social-media-graphics
claude mcp add --transport http wireflow https://www.wireflow.ai/api/mcp
```

To check the setup, ask for a post. The skill first looks for the Wireflow tools `run_app`, `get_execution`, `get_credit_balance` and `whoami`. If they are missing, it says so in one line, gives you the setup steps, hands over a filled-in post sheet and design spec, and stops there. The [agent skills](/docs/skill) page covers the install options for every Wireflow skill.

## What do I tell Claude to make a quote card, stat post or announcement?

Tell Claude what the post says and which brand it is for, in plain words:

```text
Make an announcement post for Brightline Content: "We just opened our Austin studio".
Open house, Saturday 14 November, 5 to 8 pm. Warm off-white background (#FAF7F2),
charcoal text (#222222), coral accent (#FF6B4A), bold rounded sans serif.
```

The skill asks three questions to get the post:

1. What is the post: a quote and who said it, a stat and where it comes from, or an announcement (what, when, where)?
2. Which brand: colors, font feel, and the name or handle to show as plain text?
3. Is this one post, or a batch for a content calendar?

A quote needs a real speaker who agreed to it, and a figure needs a real source. If either is missing, the skill asks or leaves it out.

Claude turns your answers into the three inputs the app takes, and sends all three every run:

| Input | What goes in it |
| --- | --- |
| `post_text` | Every word on the graphic, each piece in double quotes. Anything outside quotes is a label and stays off the image |
| `post_type` | Quote, stat or announcement, plus how the pieces are arranged and any single accent shape |
| `brand_colors` | Background, text and accent colors (hex if known) and the font feel |

Before anything renders, the skill settles the words with you:

- It cuts hard: the hero line is 10 words or fewer and the whole post 25 or fewer, because short posts render cleanest and read on a phone.
- It keeps to three pieces at most: the hero, one supporting line and one name or handle. In Wireflow's test, the smallest piece was the one that went missing, so that last piece is made large enough to matter, or dropped.
- If the post has a date, it checks that the weekday and the date agree, because the app draws whatever it receives.
- It shows you the exact text for the graphic, and you can change any of it. Afterwards, a word change means a new run.

## Can I put my logo or a photo on the graphic?

No, the app takes text only, so a logo file or a photo cannot go in, and its design is fixed as a flat graphic with no photos of people and no logos. Your brand name or handle can still go on the post as plain text.

The app adds no words of its own, and MagicPrompt is off, so nothing rewrites your copy. The skill's guardrails also rule out third-party logos or trademarks shown as endorsements.

## What does one run give back, and where are the files saved?

One run gives back three PNGs of the same post, and the skill saves them under `./wireflow-outputs/`:

| Size | Pixels | Ratio |
| --- | --- | --- |
| square | 2048x2048 | 1:1 |
| portrait | 1728x2304 | 3:4 |
| story | 1440x2560 | 9:16 |

Each file goes into a dated `social-graphics` folder and is named after the post and its size, for example `./wireflow-outputs/YYYY-MM-DD/social-graphics/post-slug_story.png`.

While the run is `PENDING`, `RUNNING` or `IN_PROGRESS`, the skill polls `get_execution` every 5 seconds. In Wireflow's test, a run took about a minute. On `FAILED`, `CANCELLED` or `TIMEOUT`, it stops and quotes the error.

It closes with a table of each size, its file and its link, plus the credits spent (your balance before minus after). Then it offers the next post in your calendar, and the skill's board in Wireflow for adding sizes or a fixed brand look.

## Can it make a 4:5 post for the Instagram feed?

Not directly: Ideogram 4.5 has no 4:5 size, so the portrait is 3:4, which Instagram and LinkedIn both accept. The app keeps the words out of the top and bottom 10 percent, so a designer or scheduler can crop the portrait to 4:5 without losing text.

The app always draws the same three sizes. For other sizes, use the skill's board in Wireflow, which the skill offers at the end of a run.

## Will the words on the graphic come out exactly as I wrote them?

Not always, so the skill opens all three images and checks each one against the text you approved. The sizes are drawn separately, so one can be wrong while the others are right. It checks five things:

- Every quoted piece is there. A dropped small line is the most likely miss: in Wireflow's test, the square lost its footer.
- Each word, number, capital and punctuation mark matches letter by letter, and nothing was added.
- The type weight stays even within a line, and letters from one line do not touch the line below.
- The brand colors are right. If the layout differs between sizes, such as centered in one and left aligned in another, the skill tells you.
- On the story, the words stay clear of the top and bottom app buttons.

It lists each miss per file. A fix is a new run of all three sizes for 57 credits, after your yes, and the skill never calls a post ready to schedule while a miss is open.

## How many credits does one social media graphic cost?

One run costs 57 credits for the post in all three sizes, which is 19 credits per image (Ideogram 4.5, measured on 2 November 2026).

| What | Credits |
| --- | --- |
| Settling the words and quoting the cost | 0 |
| One post in square, portrait and story sizes | 57 credits per run, 19 per image |
| One size only | 57, because the app always draws all three |
| A fix after the check | 57 more, only after your yes |

Before it runs, the skill calls `whoami` and `get_credit_balance` and quotes the total with your balance, for example: "1 post x 57 credits = 57 credits for 3 sizes." For a batch, it multiplies and lists the posts, then makes one run per approved post. If your balance is short, it stops and says so. Nothing renders until you say yes to these words and this price.

Rendering needs credits on your Wireflow account. See [pricing](https://www.wireflow.ai/pricing) for what each plan includes.

## Does the skill work in Cursor, Codex or Claude Desktop?

Yes, the skills CLI installs it in Cursor, Codex, Gemini CLI and other agents that read `SKILL.md`, and the skill is also written for Claude.ai and Claude Desktop. Each agent must be connected to the Wireflow MCP server before the skill can render anything. The skills are built for Claude Code, and Wireflow notes that it has not tested them 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, and the report at the end of a run lists a link for each size.

For connection steps, see the guides for [Cursor](/docs/generate-images-and-videos-in-cursor) and [Codex](/docs/generate-images-and-videos-with-codex), or [Use Wireflow from your agent](/docs/use-wireflow-from-your-agent) for every client, including ChatGPT. To run your own Wireflow workflows from code with an API key instead, see the [run API](/docs/api/run).

Last updated: 2026-11-14

---

Documentation index: fetch https://www.wireflow.ai/llms.txt for the full list of Wireflow docs.
