SillyTavern Character Cards Explained: V2 vs. V3 Specifications

---

Most people think downloading a PNG from a character hub means getting a 2D anime girl. But if you've spent any time in the local LLM or SillyTavern scene, you know that 2MB file is actually a compressed brain.

Inside that image is a meticulously crafted 10,000-token backstory, specific dialogue examples, and an entire world state.

Legacy V2 cards did the job when we were struggling with 2k context limits. But with 32k and 128k windows becoming the norm, forcing complex lore into a single description field is a massive waste of tokens. V3 is the modern standard, built to handle alternate scenarios, group chats, and deep lore without breaking the prompt structure.

Here is exactly how it works under the hood, and why you need to upgrade.

---

1. How Does a Picture Hold a Prompt? (PNG Steganography)

If you've ever wondered how a literal picture contains a 10k-token prompt, the answer is PNG Steganography.

PNG files are split into different "chunks." While most chunks handle the pixel data you see on your screen, the format also allows for custom tEXt chunks. Creators use these specific text chunks to shove a raw, Base64-encoded JSON payload right into the image.

When you drop that card into a frontend like SillyTavern, it completely ignores the image pixels. It unpacks the JSON and feeds the prompt directly to your LLaMA 3 or OpenRouter API. It’s portable and brilliant.

---

2. V2 vs. V3: What Actually Changed?

The original V2 spec gave us the basics: name, description, personality, and mes_example. It was fine for simple one-on-one chats. But as creators started building RPG scenarios and interconnected lore, they hit a wall.

V3 fixes this by wrapping the character logic inside a dedicated data block and opening up the extensions dictionary.

Feature Comparison

The V3 JSON Structure: Under the Hood

Unlike V2, which dumps everything at the root level, V3 is structured for token efficiency. Frontends can now selectively load elements instead of dumping the entire file into the context window.

{
  "spec": "chara_card_v3",
  "spec_version": "3.0",
  "data": {
    "name": "Aria",
    "description": "...",
    "alternate_greetings": [
      "You find Aria sitting alone in the tavern...",
      "Aria draws her sword the moment you enter the room."
    ],
    "character_version": "1.2",
    "extensions": {
      "depth_prompt": {
        "prompt": "Remember to speak in archaic English.",
        "depth": 4
      }
    }
  }
}

---

3. Stop Corrupting Your Cards

You could extract the JSON, edit it in VS Code, and inject it back via a Python script. But nobody actually wants to do that every time they tweak a prompt.

Most people use web editors. But if you edit the metadata with a sloppy tool—or worse, open the PNG in Photoshop and hit save—you will wipe the text chunks entirely. Your character gets lobotomized.

When you start writing highly specific Regex rules or deep Lorebooks, you need a robust editor that explicitly supports V3 extensions without stripping them out upon save.

---

4. The Local Dilemma: Dead Drives vs. Cloud Censorship

If you are a strict local user, you face a massive dilemma. You keep your V3 PNGs strictly on your local SSD to avoid prying eyes. But the second your drive dies, or you format your PC without a backup, boom—hundreds of hours of carefully crafted lore, custom regex rules, and fine-tuned personas are permanently nuked.

So, what's the alternative? Syncing your cards via standard cloud drives (Google Drive, Dropbox) or using centralized chat apps? If you are running unfiltered, NSFW, or highly private scenarios, uploading your plaintext JSON to corporate servers is begging for automated bans, silent censorship, or data leaks.

You shouldn't have to choose between losing your data and compromising your privacy. This is exactly why we built Abolitus.

Abolitus gives you the convenience of PC-to-Mobile cross-play using a Zero-Knowledge E2EE (End-to-End Encryption) Cloud Sync, while maintaining 100% SillyTavern (ST) compatibility.

• AES-256-GCM Encryption: Before any character card, lorebook, or chat log leaves your device, it is encrypted locally. The encryption key never leaves your hardware.
• Blind Cloud Sync: When you sync your data to your phone, our servers only receive unreadable, encrypted blobs. We cannot scan your images, read your lore, or censor your chats. We literally do not have the keys.
• Flawless ST Ecosystem Support: You don't have to abandon your workflow. Abolitus natively supports ST V2 and V3 formats. You can drop in your existing ST cards, sync your encrypted tavern to the cloud, and pick up the roleplay on your phone—without friction and without fear.

You get to keep your data backed up and synced, while maintaining the absolute privacy of a purely local setup.

---

5. FAQ

Q: Will my old V2 cards stop working? No. Almost all frontends, including Abolitus, maintain backward compatibility with V2. But if you export or save edits, you should upgrade them to V3 to unlock alternate greetings and extensions.

Q: Does W++ formatting work better in V2 or V3? W++ works the exact same across both versions since it's just a writing style placed inside the description field. That said, there is a massive debate in the community right now regarding W++ vs. Natural Language—which we’ll break down in our next post.

Ready to chat without anyone looking over your shoulder? Import your V3 cards into Abolitus today and experience true, encrypted, uncensored roleplay.