📖 Story Loom — Roleplay-to-Novel Converter

Story Loom is a content-agnostic story-weaving engine that transforms fragmented character-turn roleplay into unified, publishable prose. Feed it the choppy output of any scenario—multiple character voices, first-person protagonist turns, disjointed scene beats—and receive back a single cohesive narrative that reads like a novel, not a chat log.

What It Does

• Merges multi-voice character turns into seamless story prose
• Converts first-person protagonist content into third-person close perspective
• Filters out meta-directive commentary (narrative instructions disguised as character thoughts)
• Transforms exposition into shown experience (show-don't-tell enforcement)
• Elevates vague descriptions into specific, lore-accurate worldbuilding
• Applies editorial discretion: rewrites when needed, preserves when it serves the story

How to Use

Step 1: Configure the Engine

• Copy the SETTING section from the scenario you want to convert
• Paste it into Story Loom's <SETTING> block (replace the template)
• Copy the STYLE section from the scenario you want to convert
• Paste it into Story Loom's <STYLE> block (keep the <weaving_directive> block—it's universal)

Step 2: Import Characters & Locations

• Add all characters from the original scenario into Story Loom's character slots
• Add the {{user}} character as a named character (the protagonist the engine references for first-to-third person conversion if required)
• Import any locations/objects from the original scenario

Step 3: Feed It Content

• Copy up to 2500 tokens of character interaction at a time
• Paste into your message with an instruction like: "Unify this section into a whole"
• The engine will transform the fragmented turns into cohesive prose

Step 4: Iterate

• Continue feeding sections of 2500 tokens or less
• The engine maintains narrative flow across sections
• Add steering instructions if needed: "Prioritize [Character]'s perspective" or "Emphasize the tension in this scene"

Tips

• Smaller chunks (1500-2000 tokens) often produce better results than pushing the limit
• If a scene feels rushed, try feeding it again with: "Slow this down, add more sensory detail"
• The engine will preserve dialogue meaning but may rewrite surrounding prose for cohesion—this is intentional
• Meta-directive content (characters announcing what "will" happen) is automatically filtered—trust the process