X (Twitter) and Threads post images
X and Threads are text-first platforms. Images aren't there for aesthetics — they're there to reinforce the argument and boost reach.
Real-world numbers: posts with images vs. text-only posts get:
- 2.3x more impressions
- 2.8x more reposts
- 1.9x more replies
But the goal here flips compared to Instagram — on X, an image is an "information weapon", not decoration. Whether your readers can grab your point in half a second decides whether they keep reading the post.
This chapter covers the full prompt templates for the three core image types on X and Threads (infographics / quote cards / data charts), plus visual playbooks for tech influencers vs. founders vs. investors.
1. Key Dimensions
| Platform | Ratio | Recommended Pixels | Notes |
|---|---|---|---|
| X (Twitter) | 1.91:1 | 1200×628 | Timeline preview |
| X carousel | 1:1 | 1080×1080 | Image carousel |
| Threads | 1:1 / 1.91:1 | 1080×1080 / 1200×628 | Same as X |
| Link preview | 1.91:1 | 1200×628 | Same as LinkedIn article cover |
The 1.91:1 ratio for the X timeline is the golden size — anything else gets cropped or shows black bars.
2. Infographic Prompt Template (Data / Comparison)
Twitter card, 1.91:1, 1200×628.
Top large title (sans-serif, dark navy):
"Junior dev salary benchmarks 2026"
Three columns with horizontal bars (left-aligned):
- Left column: "USA" — bar showing "$95K" (in dark teal)
- Middle column: "AU" — bar showing "$78K" (in dark teal)
- Right column: "EU" — bar showing "€55K" (in dark teal)
Each bar:
- Country code on left (small all-caps gray)
- Bar with numeric label inside (white text on teal bar)
- Length proportional to value
Bottom-right small attribution (italic light gray):
"@yourhandle · Source: Glassdoor 2026"
Top-left small bracket text (gray):
"[ thread 1/8 ]"
Style: minimal flat infographic, off-white background,
single accent color (dark teal), readable from mobile,
clean editorial.
Exact text only, no extra elements, no decorative shapes.
Three Iron Rules for Infographics
- Data bigger than text — "$95K" should be 3x the size of "USA"
- Single accent color — one main color per image, no rainbows
- Bottom signature + data source — boosts credibility, lifts repost rate by 40%
3. Quote Card Prompt Template
Twitter card, 1.91:1, 1200×628.
Center large bold quote (Helvetica or similar geometric sans-serif,
deep navy text, on off-white background):
"The hardest part of AI engineering
isn't the AI. It's the engineering."
Bottom-right small italic (gray):
"— @yourhandle"
Top-left small bracket text (light gray):
"[ thread 1/8 ]"
Style: minimal quote card, off-white background,
serif quote text in deep navy,
generous margins (15% padding from edges),
clean editorial card.
NO decorative quotes (no “” glyphs).
NO color accent — pure off-white + dark navy.
Exact text only.
Why Quote Cards Work
The fastest-spreading content on X is the counter-intuitive claim:
- ❌ "AI engineering is hard" — boring
- ✅ "The hardest part of AI engineering isn't the AI. It's the engineering." — flip + memorable
Quote cards turn that kind of claim into a visual weapon — readers screenshot and forward, spreading 3x faster than plain text.
4. Three Account Visual Playbooks
There are three types of high-follower accounts on X, and each plays a different visual game:
Type A: Tech Influencers (@dan_abramov / @swyx / @pcottle style)
Visual weapons: code screenshots / architecture diagrams / benchmark charts.
Twitter card, 1.91:1, 1200×628.
Subject: a code editor screenshot showing TypeScript code,
syntax highlighted (VS Code dark+ theme),
specific function definition with type annotations.
Top label (small all-caps gray):
"TYPESCRIPT TIP"
Bottom annotation (small white):
"This pattern saves 40% bundle size"
Style: developer aesthetic, dark theme,
crisp readable code (no cursive fonts),
authentic IDE feel.
Tone: editorial, restrained, reader-focused. Palette: dark background + 1 accent (green / yellow / blue).
Type B: Founders (@levelsio / @paulg style)
Visual weapons: revenue charts / product screenshots / growth graphs / real workspace shots.
Twitter card, 1.91:1, 1200×628.
Subject: a graph showing monthly recurring revenue
growing from $0 to $30K over 12 months,
warm hand-drawn aesthetic (not corporate stiff),
slight imperfections feel authentic.
Top text (medium bold dark navy):
"Year 1 of @nomadlist"
Bottom text (small italic):
"$0 → $30K MRR · solo founder · no funding"
Style: personal authentic aesthetic, warm tones,
slightly hand-drawn / sketchy edges,
"in public" vibe.
Tone: personal, warm, real. Palette: warm tones, brand colors.
Type C: Investors / Thought Leaders (@balajis / @naval style)
Visual weapons: quote cards / concept diagrams / thinking frameworks.
Twitter card, 1.91:1, 1200×628.
Subject: a conceptual 2x2 matrix diagram,
labeled axes:
- Horizontal: "Short-term ↔ Long-term"
- Vertical: "Predictable ↔ Uncertain"
- Each quadrant labeled with one word + one example
Style: minimalist consulting deck aesthetic,
2×2 matrix in light gray lines,
crisp typography,
intellectual feel.
Tone: thinking / framework-driven / discussion-starting. Palette: white + black/gray, max 1 accent color.
5. Threads vs X — The Subtle Differences
Threads and X are 90% the same, but the 10% that's different decides what content lands:
| Dimension | X (Twitter) | Threads |
|---|---|---|
| Text length | 280 chars | 500 chars (and longer) |
| Image tone | Information-dense / data-heavy | Editorial / lifestyle |
| Algorithm preference | Strong boost for image posts | Friendly to text-only too |
| User demographic | Tech / media / VC | Meta users / lifestyle |
| Sharing culture | Repost / quote tweet heavy | Reply / discussion heavy |
In practice: make two image versions of the same content — data/quote cards for X, softer editorial style for Threads.
6. Things That Went Wrong
Fail 1: Can't Read on Mobile
The infographic looked fine at 1200×628 on desktop, but X is mostly mobile — shrunk to 600×314, the text turned to mush.
Fix: infographic text must start at 24pt minimum (more than 1/12 of the image height). Data numbers should be at least 60pt.
Fail 2: Over 4MB
X aggressively compresses files over 4MB, blurring your text.
Fix: run it through TinyPNG before upload, get it under 2MB. Complex backgrounds, gradients, and noise inflate file size — simplify wherever you can.
Fail 3: Too Many Emojis
Tweet card + caption with five emojis — looked unprofessional and got me unfollowed by the serious crowd.
Fix: tweet card ≤ 1 emoji (or zero), caption ≤ 2 emojis. Serious X users are sensitive to emoji spam.
Fail 4: Brand Colors All Over the Place
Different palette every post — four weeks later my profile looked like four different accounts.
Fix: lock in 1-2 accent colors (something like deep teal + warm orange) and bake them into every prompt.
7. What We Learned
We coached an AI Engineer doing content on X for 4 weeks (baseline 3K followers). After switching prompt strategy:
| Stage | Posts/month × avg images | Avg impressions | Monthly follower growth |
|---|---|---|---|
| Before (no images / random images) | 20 × 0.3 | 1,200 | +20 |
| With images (infographics + quote cards) | 20 × 1 | 8,500 | +340 |
What the data shows: posts with images are the baseline traffic floor, infographics are the amplifier. Every post with an image = 7x reach.
8. What's Next
The next chapter, Ch 17: Failures and the Post-Production Combo, wraps up this whole direction — systematizing the four classic failure modes you'll hit on every platform (hands / text / copyright / NSFW) plus the Photopea / Figma post-production rescue workflow.
If you're in a rush to ship one image post right now:
- Pick which account type you're playing (tech influencer / founder / investor)
- Grab the matching template (§4)
- Choose infographic (§2) or quote card (§3)
- Generate 4 versions, pick 1
- Compress to under 2MB before uploading
The real win condition for X / Threads images is the "screenshot-share rate" — whether your image makes readers screenshot it and DM "check this out" to a friend decides how high your virality coefficient goes.
📷 High-Repost X / Threads Image Examples
From awesome-gpt-image (CC BY 4.0). Two real cases that show what kind of images go viral on X.
Case 1: 100 Tech Topics Mega Grid (Information Weapon)
Prompt (excerpt):
Create a 10 × 10 grid of 100 different topics representing recent technological progress.
Use a realistic, polished editorial illustration style.
Each topic should appear in its own square with a short clear label underneath.
Use these row themes: Row 1: AI models and agents, Row 2: robotics,
Row 3: semiconductors and compute, Row 4: networks and smart devices,
Row 5: biotech and health technology...
100 different tech topics generated in one shot, each with the right text label. This kind of "extreme information density" image gets crazy repost rates on X — readers screenshot and DM it to peers saying "whole industry in one image". It's the "deluxe edition" of the §4 infographic template, the format tech influencers and investors use most.
📷 Creator: @chetaslua · Curated by: awesome-gpt-image
Case 2: Codex Chalkboard Article Visualization (Premium Quote Card)
Prompt:
A picture says a thousand words. GPT Image 2 creates them.
An 8-word prompt produces a chalkboard-style, long-document visualization with serious density. This is the canonical viral pattern of "one-line post + one visualization image". When you're posting a long-form thinking thread on X, dropping in a chalkboard-style image like this gets 2-3x the repost rate of a regular image.
📷 Creator: @gabrielchua · Curated by: awesome-gpt-image
❓ 常见问题
关于本章主题最常被搜索的问题,点击展开答案
X (Twitter) 推文卡片尺寸?
1.91:1(1200×628)。多图轮播用 1:1(1080×1080)。X 时间轴预览图严格 1.91:1,其他比例会被裁切。
X 配图为什么要用?
带图推文比纯文字曝光率高 2.3 倍、转发率高 2.8 倍、评论率高 1.9 倍(实测数据)。但配图目的是"强化论点",不是装饰——0.5 秒看图能不能 get 到论点决定继续读不读。
X vs Threads 配图差异?
X 信息密 / 数据多 / 图片推文算法 boost 强;Threads 更 editorial / 生活化 / 文字推文也友好。同一内容做两版——X 用数据图 / 引言图,Threads 用更柔的 editorial 风格。