1. Upload Voxta character.json

2. Select a Portrait

Voxta keeps images separate. Upload the character's face here to embed it into the card.

Technical Mapping: From Voice to Text

Voxta uses a unique schema optimized for TTS contexts. Converting it back to a standard roleplay card requires specific mapping logic:

1. Variable Normalization

Voxta often uses {{Char}} and {{User}} (CamelCase variables). Standard TavernAI expects {{char}} and {{user}} (lowercase). Our tool scans the dialogue examples and greeting for these variables and standardizes them, ensuring the character recognizes the user immediately in SillyTavern.

2. Field Re-Alignment

We map the Voxta fields back to their V2 equivalents:

  • firstMessage → first_mes
  • personality → personality
  • profile → description (or concatenated with personality)
  • creatorNotes → creator_notes (Tavern Extension)

3. Base64 Injection

Finally, we take this normalized data, encode it into a UTF-8 Base64 string, and inject it into the tEXt chunk of the PNG you uploaded. This creates a fully valid V2 card.

Conversion Guide

  1. Locate JSON: Go to your Voxta `profiles` folder and find the `character.json` for your bot.
  2. Upload: Drag that file into the tool above.
  3. Add Image: Grab the `character.png` from the same folder and upload it in Step 2.
  4. Convert: Generate your new, portable PNG card.

FAQ: Voxta Exporting

Q: Will the voice settings be saved?

A: No. TavernAI cards do not support ElevenLabs or Coqui settings. You are exporting the *personality*, not the *voice*.

Q: Can I put this card back into Voxta later?

A: Yes! Use our "Any to Voxta" tool to convert it back if you ever lose the original profile.