HoneyChat HoneyChat
HoneyChat ·From $4.99/mo · Free: 20 msg/day · No signup See plans →

How to Put Images in JanitorAI First Message — Setup Guide (2026)

· · David Mercer · 4 min read
How to Put Images in JanitorAI First Message — Setup Guide (2026)

Short answer: Use markdown syntax ![alt text](image_url). Upload your image to Imgur or Catbox first, then paste the direct URL. JanitorAI doesn’t host images natively. For AI-generated images per scene without uploading anything, HoneyChat has built-in per-character LoRA image generation — 3 photos/day free, 30/day on Premium ($9.99/mo).

Want AI images that generate automatically in chat, no upload, consistent character face? Open below.

Native AI image generation in HoneyChat

Quick options

  • Want to embed pre-uploaded images in JanitorAI → Use markdown + Imgur/Catbox URL
  • Want AI to generate images per scene automaticallyHoneyChat ($4.99/mo, native gen)
  • Want polished web with photo + AI videoCandy AI ($12.99/mo)
  • Want NSFW image gen + companionOurDream AI ($9.99/mo)

JanitorAI is a character chat platform, not an image generator. The platform’s design is “you bring the LLM (BYOK), we provide the character infrastructure.” Images in JanitorAI are pre-uploaded URLs you embed using markdown syntax — they’re static, not generated.

If you want AI-generated images that change per scene (your character sending a selfie after you chat about something), JanitorAI doesn’t do that natively. Let me walk through both paths.

Step-by-Step: Embed an Image in JanitorAI

1

Prepare your image

Have a .jpg, .png, .webp, or .gif file under 5MB. Higher quality = better but slower to load.

2

Upload to image host

Best options: imgur.com (free, reliable), catbox.moe (no signup, NSFW-friendly), postimg.cc. Avoid Discord CDN (URLs expire after a while).

3

Get the DIRECT image URL

After upload, copy the URL ending in .jpg/.png/.webp. Test by pasting in new tab — should show only image, not a webpage.

4

Verify HTTPS

URL must start with https:// — JanitorAI rejects HTTP for security. Most image hosts default to HTTPS now.

5

Add markdown to JanitorAI

In character first message field: ![character introduction](https://i.imgur.com/yourimage.jpg). Save the character.

What Image Hosting Works (and What Doesn’t)

Yes Imgur (most reliable)
Yes Catbox.moe (NSFW-friendly)
No Discord CDN (URLs expire)
5 MB Practical file size limit

Works reliably:

  • Imgur — free, fast, direct links via i.imgur.com/[hash].jpg. Account optional. Most reliable.
  • Catbox.moe — no signup, drag-and-drop, NSFW-friendly. URLs via files.catbox.moe/[hash].jpg.
  • Postimages — free with optional account. URLs via i.postimg.cc/[path].jpg.
  • ImgBB — free, embed-friendly. URLs via i.ibb.co/[path].jpg.

Doesn’t work well:

  • Discord CDN — URLs from Discord attachments expire (CDN tokens rotate). Image breaks within days.
  • Google Drive direct links — drive.google.com URLs require auth, don’t render in JanitorAI.
  • Dropbox direct links — similar issue, mostly broken.
  • Twitter/X image URLs — heavy hotlink protection.

For NSFW first messages, Catbox.moe is the standard choice. Imgur’s policy on NSFW is complicated — sometimes works, sometimes flagged.

Markdown Syntax Examples

Single image in first message:

*She walks into the room and smiles at you.*
![Character introduction pose](https://i.imgur.com/yourimage.jpg)
"Hey, I was wondering when you'd show up."

Multiple images (introduction collage):

![Full body](https://files.catbox.moe/pose1.png)
![Close-up](https://files.catbox.moe/pose2.png)
![Outfit detail](https://files.catbox.moe/pose3.png)

Image with alt text for accessibility:

![Anna sitting on a windowsill in afternoon light, smiling at camera](https://i.imgur.com/anna01.jpg)

With surrounding narrative:

The door swings open and Sarah peeks her head in.
![Sarah peeking through doorway, mischievous grin](https://files.catbox.moe/sarah-door.jpg)
"Oh, you're awake! I was just about to wake you up."

Why Images Don’t Load (Troubleshooting)

If your embedded image doesn’t render in JanitorAI:

1. URL is not direct

Wrong: https://imgur.com/gallery/abc123 (page, not image) Right: https://i.imgur.com/abc123.jpg (direct file)

Fix: open the original image in a new tab on the host site, right-click “Copy image address”, use that URL.

2. URL is HTTP not HTTPS

Wrong: http://example.com/image.jpg Right: https://example.com/image.jpg

Fix: change http:// to https://. If the host doesn’t support HTTPS, switch hosts.

3. Hotlink blocked

Some image hosts (Pinterest, Instagram, Twitter) block external embedding. The image loads in their site but not when JanitorAI tries to render it.

Fix: download the image, re-upload to Imgur or Catbox, use new URL.

4. File too large

Practical limit is around 5MB. Larger files may not render or load extremely slowly.

Fix: compress the image (tinypng.com, squoosh.app), then re-upload.

5. Markdown syntax error

Wrong: ![ alt ] ( url ) (spaces around brackets) Right: ![alt](url) (no spaces)

Wrong: ! [alt](url) (space between ! and [) Right: ![alt](url)

Fix: ensure no spaces between the markdown symbols.

6. NSFW content auto-removed by host

Imgur sometimes auto-removes NSFW after upload. The URL initially works, then breaks days later.

Fix: use Catbox.moe for NSFW images — more permissive policy.

When You Want AI-Generated Images Instead

The fundamental limitation of JanitorAI’s image approach: images are static. You upload one version, the character shows that one version every time. There’s no “the character sends you a new selfie based on what we just discussed.”

For dynamic per-scene image generation:

HoneyChat generates images in chat using per-character LoRA. Each generation is a new image, but the character’s face and body stay consistent because of LoRA training. NSFW images are tier-gated by content level (0-5). 3 photos/day free, scaling up to 150/day on Elite tier.

Candy AI generates photos tied to character presets. Quality is high, less LoRA-consistent than HoneyChat but still good. Web only, $12.99/mo base.

OurDream AI similar — character-tied photo gen at $9.99/mo.

The difference from JanitorAI is that on these platforms, images are part of the conversation flow rather than static embeds. Your AI girlfriend can “send a selfie” that’s actually a fresh generation matching context.

JanitorAI image embed vs native AI image gen

JanitorAI + Markdown HoneyChat Candy AI OurDream AI
Image type Static embed Generated per scene Generated per scene Generated per scene
Character consistency Same image every time Per-character LoRA Preset-based Preset-based
Setup Upload + markdown None (built-in) None (built-in) None (built-in)
Scene-context aware No Yes (LLM prompts gen) Yes (limited) Yes (limited)
Daily image limits Unlimited embed 3 free, 30 Premium 20 Base, unlimited Premium Plan-dependent
NSFW images Pre-uploaded only Tier-gated 0-5 levels Yes on paid Yes on paid
Cost Free (BYOK LLM cost) $4.99-$39.99/mo $12.99-$39.99/mo $9.99/mo+

If you want AI image gen automatically in chat — no upload, no markdown, no static images — start below.

Pre-Generated Image Strategies for JanitorAI

If you’re staying on JanitorAI and using static embeds, here are the strategies users employ:

1. Multi-pose character intro — embed 2-4 images at the first message showing different angles or expressions of your character. Establishes visual identity strongly.

2. Scenario-themed images — for character configs with specific roleplay scenarios (vacation, gym, work), pre-upload images matching the scenario and embed in first message.

3. Character + setting combo — wide-angle landscape image of the setting alongside the character close-up. Establishes location.

4. Outfit variations — for long roleplay arcs, upload variants of the character in different outfits/situations. Edit the character config periodically to swap which image shows in first message.

5. NSFW preview gating — start with safe-for-public first message image, escalate to NSFW after consent in conversation. Pre-upload both, edit character to swap as scene progresses.

These all work within JanitorAI’s static-embed model. They don’t replicate the dynamic generation of HoneyChat or Candy AI.

What I’d Recommend

If you want to stay on JanitorAI: Catbox.moe + markdown syntax for the most reliable image embedding, including NSFW. Standard markdown ![alt](url).

If you want dynamic AI image gen in chat: HoneyChat Basic ($4.99/mo, 10 photos/day) or Premium ($9.99/mo, 30/day). Free tier (3 photos/day) lets you evaluate.

If you want polished photo + AI video: Candy AI at $12.99/mo Base or $24.99 Premium for unlimited photos + video.

If image gen is the only reason you’re considering switching from JanitorAI: try HoneyChat’s free tier for a day. You’ll see whether dynamic generation matters to you. If it doesn’t, stick with JanitorAI + markdown.

I covered the broader JanitorAI setup in JanitorAI API key with OpenRouter, the LLM selection in Best free LLM for JanitorAI, and the voice option in How to have JanitorAI speak to me.

Sources & References

FAQ

Can JanitorAI display images in the first message?

Yes. JanitorAI supports markdown image syntax in character first messages and chat responses. Use `![alt text](image_url)` format with a publicly accessible image URL. The image will render inline in the chat. JanitorAI doesn't host images — you need to upload to Imgur, Catbox, or similar and use that URL.

What image hosting works with JanitorAI?

Imgur (free, reliable, direct image links), Catbox.moe (no signup, NSFW-friendly), Postimages, ImgBB. Avoid Discord CDN URLs — they expire and break embedded images. Use direct image links ending in .jpg, .png, .webp, or .gif. URLs must be HTTPS for security policies.

Why are images not loading in my JanitorAI character?

Common causes: (1) image URL is not direct — should end in .jpg/.png; (2) hosting service blocks hotlinking — try Catbox.moe instead; (3) URL is HTTP not HTTPS — JanitorAI requires HTTPS; (4) markdown syntax error — use `![alt](url)` with no spaces; (5) image is too large — keep under 5MB. JanitorAI doesn't natively generate images — for AI-generated images use HoneyChat or Candy AI.

Does JanitorAI generate images natively?

No. JanitorAI is a character chat platform with text-only output. Images you embed are pre-uploaded URLs you control. For native AI image generation in companion context, HoneyChat (per-character LoRA art, 30+ images/day on Premium), Candy AI (polished photo gen), or OurDream AI all generate images automatically as part of the chat experience.

Can I include multiple images in JanitorAI first message?

Yes, JanitorAI markdown supports multiple `![alt](url)` blocks in the same message. Each renders inline. For character introduction with multiple poses, you can include 3-5 images. Too many slows down message rendering. For dynamic generated images per scene, HoneyChat's native image gen creates new images each conversation rather than relying on pre-uploaded URLs.

Related Articles

Ready to Meet Your Companion?

Free: 20 messages/day. Premium starts at $4.99/mo.

Chat in Browser Telegram Bot