Team

A lorebook is a collection of entries injected into your story context. Always active entries are included on every request.

Dynamic entries are keyword-triggered: when their keywords appear in recent messages they're pulled in automatically, letting you build out "unlimited" lore with only what's relevant surfacing at a time. Up to 4 dynamic entries share a token budget per request (500 tokens for Basic, 1,500 for Pro) and may be cut to fit. If an entry is long, only the beginning is included, so keep the most important info at the top. For details you always want the AI to know, use Always active instead.

Where it applies: attach a lorebook to a project and every story in that project uses it. Attach one to a single story for story-specific lore. If both exist, the story's lorebook takes precedence when its Override toggle is on; otherwise the story falls back to the project's.

Always-active limit: enabled always-active entries may use up to 8,000 characters or 2,500 tokens per lorebook, whichever allows more text, with a 100,000-character safety limit. Entries are included top-to-bottom until the limit is reached. Anything past the cutoff is skipped. Put your most important entries at the top. Disabled entries don't count, so you can stash overflow lore and toggle entries in and out as needed. Keep what's enabled short and stable - facts the model always needs (character traits, world rules, recurring places). For big lore dumps, dynamic entries or file uploads are usually a better fit.