Smart Captions

Generate AI-powered captions for your images with themed prompts and full style control

Smart Captions lets you upload or fetch an image and instantly generate a themed AI caption overlaid on it. Choose from five built-in caption themes, write your own custom prompt, then fine-tune the text appearance before downloading or saving to your Media Vault.

How It Works

The Smart Captions workspace is a three-panel layout:

• Left panel — Image source, caption theme, and text style controls
• Center panel — Live canvas preview showing your image with the generated caption
• Right panel — Creation history rail so you can revisit or remix past results

On mobile, the canvas appears first with a horizontal history strip at the top and the settings panels below.

---

Getting an Image

You can bring images into Smart Captions two ways.

Upload

Drag and drop an image onto the upload zone in the left panel, click Browse to pick a file, or paste an image from your clipboard (Ctrl/Cmd + V).

Accepted formats: JPEG, PNG, WebP. Maximum file size is 100 MB (files over 20 MB are automatically compressed). GIF files are not supported.

Reddit

Switch to the Reddit tab to pull a random image from one or more subreddits. Two input modes are available:

• Subreddit — Add up to 25 subreddits using the chip input. Joip fetches recent image posts and picks one at random.
• Search — Enter free-text keywords (e.g. “amateur selfies”) to search across Reddit for matching image posts.

Click Fetch Image (or Search Reddit) to load an image. Press Get Another to cycle to a different result without re-entering your subreddits. Your subreddit list is saved locally so it persists between visits.

You can also send images to Smart Captions from other Joip tools using the Send to… dropdown that appears on results in BabeCock Studio, Censor Studio, and more.

---

Choosing a Caption Theme

The Caption Theme panel offers six options. Each preset instructs the AI to adopt a specific voice, vocabulary, and psychological angle. The actual prompt engineering happens server-side so results are consistent and high quality.

JOI / Edging

A dominant “Mommy” or goddess voice that gives specific stroke commands, edge counts, denial instructions, and subspace language. Expect pet names like “pump puppet” and “drippy puppy,” along with precise stroke counts and hard-edge challenges.

Forced-Bi

A patient-then-impatient manipulator who frames gay acts through elaborate rationalizations tied to her image. Uses escalation tactics, specific scenarios, and the core hook: “It’s not gay if it’s for me.”

Bully / Simp

Pure mean-girl energy. Bored contempt, amused disgust, and casual cruelty aimed at the viewer as a hopeless simp. Covers Instagram stalking, out-of-your-league reminders, and rhetorical destruction sequences.

Pain / CBT

A sadistic perfectionist who commands specific pain acts with named implements, exact counts, durations, and intensity targets. Demands visible suffering and devotion through pain.

Cuckold

Graphic descriptions of encounters with “the bull” while relegating the viewer to spectator and cleanup duties. Heavy on physical detail, size comparisons, and post-encounter gloating.

Custom Prompt

Write your own AI instructions in the text box that appears. You can define a voice, personality, pet names, command types, example phrases, or any combination. The AI adapts to whatever level of detail you provide — from a single keyword to a multi-paragraph brief.

Maximum length: 1,000 characters.

For preset themes, the AI always speaks as the person in the image directly to the viewer. One subtle visual detail from the photo is woven in, then the caption immerses fully in the theme’s psychology.

---

Generating a Caption

1. Load an image

   Upload a file, paste from clipboard, or fetch from Reddit using the Image Source panel.

2. Pick a theme

   Select one of the five presets or choose Custom Prompt and write your own instructions.

3. Hit Generate

   Click the gold Generate Caption button at the bottom of the left panel (desktop) or the floating button at the bottom of the screen (mobile). The AI analyzes the image and produces a caption in a few seconds.

4. Review the result

   The center canvas updates in real time, showing the image with the caption rendered below (for wide/square images) or beside it (for tall/portrait images). A text preview also appears beneath the canvas so you can read longer captions comfortably.

---

Customizing Caption Appearance

Open the Text Style panel in the left sidebar to adjust how the caption looks on the canvas.

Text Color

Eight color swatches: White, Black, Gold, Red, Blue, Green, Purple, and Pink. Click any swatch to change the caption color instantly.

Font Family

Six font choices: Inter (default), Georgia, Arial Black, Impact, Trebuchet, and Verdana. The dropdown previews each font in its own typeface.

Font Size

Slider from 12 px to 72 px (default 24 px). The canvas auto-scales text to fit the available area, so this acts as a base size that the layout engine adapts.

Alignment

Toggle between Left and Center alignment for the caption text.

Effects

Two toggle switches sit below the core style controls:

Effect	Description
Background	Adds a semi-transparent dark rounded box behind the text. When enabled, an Opacity slider (0 — 100%) appears to control how opaque the backdrop is.
Shadow	Applies a soft drop shadow to the caption text for improved readability on busy images.

All style changes update the canvas preview instantly — no need to regenerate.

---

Actions on a Finished Caption

Once a caption is generated, the canvas header shows a row of action buttons:

Button	What it does
Remix	Re-generates a brand-new caption for the same image and theme. Costs credits again.
Copy	Copies the caption text to your clipboard.
Save	Saves the combined image+caption to your Media Vault as a tagged smart-caption asset.
Send to…	Opens the Joip cross-tool dropdown to pass this captioned image to BabeCock Studio, Censor Studio, or another tool.
Download	Downloads the canvas as an image file to your device. The server normalizes the format for maximum compatibility.

---

Creation History and Remix

Every caption you generate is automatically saved to the history rail on the right side of the screen (desktop) or the horizontal strip at the top (mobile). Up to 8 recent creations are kept in local storage.

1. Click a history thumbnail

   The canvas, caption text, theme selection, and all style settings are fully restored.

2. Remix with a new caption

   After restoring a history item, click Remix to generate a fresh caption for the same image. The original uncaptioned image is preserved separately so the remix never compounds captions on top of each other.

3. Start fresh

   Click the + button at the top of the history rail (or the dashed circle on mobile) to clear the canvas and start a new creation.

History items store the original raw image as a compressed snapshot, so remix works even after you close and reopen the page. If the snapshot is unavailable (very old items), the thumbnail still displays but the Remix button will be disabled.

---

Credit Cost

Each caption generation costs credits. The exact cost is set by the site administrator and displayed on the Generate button (e.g. “Generate — 10 credits”). Styling changes, downloading, copying, and saving to the vault are free.

If your balance is too low, an Insufficient Credits dialog appears with a link to purchase more credits.

Remixing (re-generating) counts as a new generation and costs credits again. Restoring a history item and just adjusting styles does not cost anything.

---

Tips for Better Captions

Use clear, well-lit images

The AI analyzes visual details in the photo to ground the caption. Higher-quality images with visible subjects produce more specific, immersive captions.

Try different themes on the same image

The same photo can yield wildly different results across themes. Experiment to find what works best for each image.

Write detailed custom prompts

When using Custom Prompt, the more detail you give (voice, pet names, vocabulary, example phrases), the closer the output matches your vision. Even a few keywords help, but a full brief produces the best results.

Remix multiple times

AI output is non-deterministic. If the first caption does not land, hit Remix a few times — you will often get a standout result within two or three tries.

---

FAQ

What image formats are supported?

JPEG, PNG, and WebP. Animated GIFs are not supported because the AI vision model requires static images.

Is there a file size limit?

The maximum upload size is 100 MB. Files between 20 MB and 100 MB trigger automatic compression. For the smoothest experience, images under 20 MB are recommended.

Can I edit the caption text manually?

Not directly on the canvas. Smart Captions is an AI-generation tool. If you want full manual control over caption text, use the Custom Session editor in Sessions, which lets you write and edit captions freely.

Why is the Remix button disabled?

Remix requires the original image file. If you restored a very old history item whose raw image snapshot was lost, the preview still shows but the file is unavailable for re-upload to the AI. Upload or fetch the image again to re-enable Remix.

Do style changes cost credits?

No. Only the initial Generate and each Remix cost credits. Changing font, color, size, alignment, background, or shadow is instant and free.

Where do saved captions go?

Clicking Save stores the combined image+caption in your Media Vault under the smart-caption category with automatic tagging. You can find it later on the Media Vault page.

Can I use Smart Captions on my phone?

Yes. The layout adapts for mobile with the canvas on top, a horizontal history strip, and a floating Generate button at the bottom of the screen.