Tavern Studio Download
Tavern Studio Guide - EN

Import Presets and JSON Assets into Tavern Studio

Presets and JSON assets are easy to collect and easy to forget. When you import them into Tavern Studio, the goal is not just to copy files. The goal is to make generation settings visible, editable, and attached to the chats that need them.

Use direct JSON import for individual assets. Use full SillyTavern migration when presets are part of a larger project move.

Who This Is For

  • SillyTavern users with saved generation presets.
  • Users moving API settings and prompt templates into Tavern Studio.
  • Writers who want separate presets for different styles.
  • Local LLM users tuning GGUF model behavior.

Core Content

A preset may include system prompt text, sampling parameters, response length, context size, and prompt ordering. Some JSON assets may also represent API configurations or related settings.

After import, review the result. Provider-specific settings, custom endpoints, reverse proxies, Azure OpenAI, Cloudflare Workers AI, account IDs, missing base URLs, or missing default models may need manual setup.

How Tavern Studio Handles It

Tavern Studio supports robust JSON import for preset-style assets and can map common OpenAI-compatible services during full migration. The app stores imported settings locally so they can be reused across chats.

For a complete SillyTavern import, choose the project root containing the data folder and use the preview step before importing.

Operation Steps

  1. Locate the preset or JSON asset.
  2. Import it through Tavern Studio's import flow.
  3. Open the preset manager.
  4. Review system prompt, sampling settings, context length, and response length.
  5. Check whether the model route or API base URL needs manual editing.
  6. Attach the preset to a test chat.
  7. Compare output before relying on it for long sessions.

FAQ

Can Tavern Studio import SillyTavern presets?

Yes. Compatible preset JSON assets can be imported into Tavern Studio.

Are API keys imported?

The full migration flow can import supported API configurations, but sensitive and custom routes should be reviewed manually.

Why does my imported preset behave differently?

Model provider, context length, prompt order, missing base URL, or unsupported fields can change behavior.

Should I keep one preset for every character?

Usually no. Reuse presets by style or model route, then pair them with different cards.

Can presets work with local models?

Yes. Presets can be used with local model routes as well as cloud API routes.

Next Step