Scriptease Lite

PROCESSING

Processing drafts…

LOCAL AI

Connect a local LLM

Bring your own model. Nothing is uploaded anywhere — requests go straight from this browser to a server running on your own machine (e.g. Ollama or an OpenAI-compatible server like LM Studio). If you use Ollama and the connection fails, you may need to set OLLAMA_ORIGINS before starting it.

Server type Base URL Model

API key (only if your server requires one) Context length Applies to Ollama (sent as num_ctx on each request). OpenAI-compatible servers like LM Studio set this when the model is loaded — check its model settings if you're hitting context-length errors. Enable thinking mode (lets reasoning models deliberate before answering — slower, sometimes sharper)