Edit prompts before generation Use function tool Use interactive mode Use multimodal captioning for portraits Extend free mode prompts (interactive/commands) Snap auto-adjusted resolutions Minimal response prompt processing Prompt generation connection profile Connection profile always used to generate the image prompt text. Leave empty to use the currently active model.

Provider Fallback Chain

Save the current backend/connection settings (source, model, sampler, dimensions, etc.) as named presets and arrange them into an ordered chain. Prompt templates, styles, custom entries, and the reference image library are not included.

Add current settings

Generate through the chain (auto-fallback) When enabled, each generation tries the chain presets top to bottom until one succeeds. Unreachable local backends (e.g. a powered-off ComfyUI machine) are skipped after about a second.

Reference Image Library

Use tagged reference images Upload images and tag them with a short name plus a description. If the active ComfyUI workflow uses the "%reference_image%" placeholder, the best-fitting image for the scene is selected automatically and injected like %user_avatar% / %char_avatar%.

Add reference image

RunPod On-Demand Pod

Lazy proxy URL

unknown

Warm up

Shut down

While a SillyTavern tab is open, a keepalive ping keeps the pod alive; it shuts down 15 minutes after the last activity either way. Red = off, orange = starting/downloading models, green = ready.

Source

SD Web UI URL

Connect

Authentication (optional) Important: run SD Web UI with the --api flag! The server must be accessible from the SillyTavern host machine.

stable-diffusion.cpp URL

Connect

The server must be accessible from the SillyTavern host machine.

DrawThings API URL

Connect

Authentication (optional) Important: run DrawThings app with HTTP API switch enabled in the UI! The server must be accessible from the SillyTavern host machine.

Hint: Save an API key in the Hugging Face (Text Completion) API settings to use it here. Model ID

Hint: Save an API key in the Chutes (Chat Completion) API settings to use it here.

Hint: Save an API key in the Electron Hub (Chat Completion) API settings to use it here.

Image Quality

Hint: Save an API key in the NanoGPT (Chat Completion) API settings to use it here.

SD.Next API URL

Connect

Authentication (optional) The server must be accessible from the SillyTavern host machine.

Hint: Save an API key in AI Horde API settings to use it here. Allow NSFW images from Horde Sanitize prompts (recommended)

Avoid spending Anlas (Opus tier)

View my Anlas

Hint: Save an API key in the NovelAI API settings to use it here.

API Key

Click to set

Will use Common API. Coding API is not supported!

Image Style

Image Quality

Duration

Server Type

ComfyUI URL

Connect

ComfyUI RunPod URL

Connect

API Key

Click to set

Important: The server must be accessible from the SillyTavern host machine.

ComfyUI Workflow

Pollinations.ai

API Key

Click to set

Enhance

API Key

Click to set

You can find your API key in the Stability AI dashboard.

Style Preset

API Key

Click to set

Prompt Upsampling

API Key

Click to set

Cloudflare Workers AI

API Key

Click to set

Hint: Account ID and API key are pulled from API connections.

API Type

Enhance

Duration (Veo)

Model

VAE

Sampling method

Scheduler

Resolution

Upscaler

Sampling steps

CFG Scale

Width

Height

Upscale by

Denoising strength

Hires steps (2nd pass)

CLIP Skip

Restore Faces Hires. Fix Karras

Use ADetailer (Face)

SMEA DYN Decrisper Variety+

Seed (-1 for random)

Style

Common prompt prefix Negative common prompt prefix

Character-specific prompt prefix Won't be used in groups. Character-specific negative prompt prefix Won't be used in groups. Shareable When checked, character-specific prompts will be saved with the character card data.

Chat Message Visibility (by source)

Uncheck to hide the extension's messages in chat prompts.

Extensions Menu Slash Command Interactive Mode Function Tool