SillyTavern Windows setup with OpenRouter

This tutorial walks through installing SillyTavern on Windows and connecting it to OpenRouter for character chats. The interface runs on the user's computer, while the demonstrated model connection uses an external API. The walkthrough does not demonstrate offline inference or a local LLM backend; the speaker mentions Ollama and LM Studio as alternatives covered in other videos.

Setup begins with checking that Git and npm are installed, choosing a folder in Command Prompt, and cloning the SillyTavern repository. The speaker then runs npm install and npm start. If the browser does not open automatically, the tutorial gives 127.0.0.1:8000 as the address.

The OpenRouter section covers creating an account and API key, pasting the key into SillyTavern, and selecting a model. The speaker notes that free endpoints have limits and shows a privacy setting that permits endpoints which may train on request data. Viewers should distinguish that setting from running the chat interface locally.

The remaining walkthrough introduces character descriptions, name substitution, and user personas. It briefly covers extensions, appearance settings, and lorebooks before explaining temperature, response length, and context limits. The speaker cautions that larger API contexts can cost more and that local models may face tighter context limits. Detailed voice, image, and memory setup is left to separate tutorials.