LoomScribe — Interactive Workspace

LoomScribe Prompt Engine

New Chat Reconnecting…

Prompt Engine

Active Scenario

Creative Writer General

General creative writing preset.

Writing Directives

Director's Note (Post-History) Injected after chat history for highest model recency attention.

Compiled Prompt Preview

Advanced Markdown Blocks

Settings

Sliding Context Window 16 msgs

Sets how many recent messages are sent to the model per turn. A bounded window keeps responses fast and saves prompt tokens, or turn it off to send the full conversation history.

Sliding Window Limit

Enabled (limits history to recent turns)

Messages to Retain

messages

2 (minimal) 16 (default) 100 (deep history)

Sliding window is OFF. The full conversation history will be sent to the model on every generation.

OpenAI Provider Dynamic Models

OpenAI API Key

Pinned / Quick Models Always shown in model picker

GLM Provider (Z.AI) Thinking & Fast Prose

GLM / Z.AI API Key

Pinned / Quick Models Always shown in model picker

DeepSeek Provider Built-in

DeepSeek API Key

OpenRouter API Key Gateway

Access hundreds of AI models from Anthropic, Meta, DeepSeek, Mistral, and more through OpenRouter's unified endpoint.

OpenRouter API Key

Get API Key

OpenRouter Models & Live Pricing

Add any model ID from OpenRouter to show in your model picker with live token pricing.

Quick suggestions:

Connect OpenAI-compatible API providers (e.g. Ollama, OpenRouter, Groq, vLLM).

Add Endpoint

Display Name

Base URL / Endpoint

API Key (Optional)

Model Identifier

Workspace Theme

Select Scenario Preset

Choose a narrative engine preset to apply system prompt & parameter defaults.

Preset Manager

Create, edit, import and export narrative prompt engine presets.

Select a preset on the left or click + New Preset to build one.

Drop a .json preset file here or browse to auto-fill.

Preset ID (lowercase_underscores)

Display Title