NotaeLibrary
Build rich, structured knowledge bases for anything — fictional universes, murder mysteries, family trees, fan rooms, and more. This guide takes you from first steps to power-user graphs.
← Go to My WorldsIn this tutorial
A World is a structured, AI-queryable knowledge base you build inside your Home Room workspace. Think of it as a living wiki — except instead of plain text pages, every entry is a typed entity (a character, a place, an event, an object) connected to other entities through relations.
You own your World. You choose what goes in, who can see it, and how detailed to make it. Worlds can be as small as a handful of characters for a murder mystery, or as large as a full canonical lore bible for a multi-book fantasy series.
Organised
Everything categorised by type — characters, places, events, objects.
Connected
Entities link to each other — a character can be at a place during an event.
AI-Queryable
Sage can answer questions directly from your World's structured data.
Creating a World takes under a minute. Here's the process:
Go to My Workspace → Worlds
Click 'New World'
Name your World
Set visibility (Public or Private)
Start adding entities
Every piece of knowledge in a World is stored as an entity. An entity has a name, a description, an optional image, a type, and a confidence level. There are five entity types, each serving a different role in your World.
People & Characters
Any person, real or fictional, living or historical. The backbone of most Worlds.
Locations & Settings
Physical or conceptual spaces — rooms, cities, planets, or even a website.
Events & Moments
Things that happened — with optional precise timestamps for gap-reasoning.
Items & Artefacts
Physical items, documents, photos, weapons, or anything with significance.
Background & Lore
Explanatory notes, rules, history, or any information that doesn't fit a person or place.
Characters are the most commonly used entity type. A Character represents any person — real, fictional, living, historical, or hypothetical.
Each Character has a name, a description (you decide how detailed to go), an optional profile image, and a confidence level (how certain you are this person exists or played the role you've assigned them).
For a murder mystery: their alibi, motive, known relationships. For a family tree: birth year, key life events, spouse, children. For a fan room: their canonical role in the story, traits, arcs. There's no required format — write what you'd want to know when you query Sage later.
Characters become powerful when you add Relations — linking them to Places ("was at"), Timeline events ("participated in"), Objects ("owns"), or other Characters ("is sibling of", "suspects", "mentored by"). See the Relations section for details.
A Place entity represents any location — a room in a house, a city, a country, a fictional world, or even a digital space like a website. Places answer the question: where did this happen?
Places are especially useful in murder mysteries (crime scenes, alibi locations, escape routes), historical research (battlefields, family homes, migration destinations), and fiction (kingdoms, taverns, portals).
Add a cover image to a Place using your Media Library — a map, floor plan, reference photo, or AI-generated image. This image appears as a thumbnail in the entity list and as the header image on the Place's detail page.
Timeline entities represent things that happened — moments, scenes, events, milestones. They're the most powerful entity type because they support precise temporal fields:
| Field | What it records | Example |
|---|---|---|
| Exact time | Precise moment — date + time | 2026-06-15 at 20:32 |
| Starts at | Beginning of a range | Suspect seen from 18:30 |
| Ends at | End of a range | Suspect seen until 19:30 |
| Approximate label | Human-readable fuzzy date | "circa summer 1962" |
| Timezone | IANA timezone for the event | Europe/Dublin |
When you add exact times to Timeline entities, Sage can reason about gaps: "Did Colonel Fairfax have time to leave the café at 19:30 and reach the library by 20:32?" Without timestamps, Sage can only tell you the sequence. With them, it can calculate whether an alibi holds.
Object entities represent physical items, documents, artefacts, or anything tangible that carries significance in your World. Objects answer: what was involved?
In a murder mystery, Objects might include the murder weapon, a cryptic letter, a missing jewel, or a set of footprints. In a family album, an Object might represent a photograph, a birth certificate, or a family heirloom.
A powerful convention: store each family photo as an Object entity with the image attached and a description of who appears in it, where it was taken, and approximately when. You can then create Character relations to link every person in the photo to that Object entity. Sage can then answer: "Show me every photo that includes Grandma Eileen."
Context entities are the catch-all for background knowledge that doesn't fit a person, place, event, or item. Think of them as lore notes.
Use Context for: the rules of a fictional magic system, a character's backstory before the story begins, a company's org chart description, a historical summary of a period, or any explanatory information that gives other entities meaning.
If the information primarily describes a single entity, put it in that entity's description. If it's shared background knowledge that multiple entities relate to, make it its own Context entity and link things to it.
Every entity has a confidence level — a 0–100 score representing how certain you are that the entity and its description are accurate. This is especially valuable in:
| Score | Label | Typical use |
|---|---|---|
| 0–25 | Speculation | A theory, an unverified claim, a fan reading |
| 26–50 | Possible | Plausible but not yet confirmed |
| 51–75 | Probable | Strong evidence supports it |
| 76–100 | Confirmed | Canonical, verified, or directly stated |
Every entity can have a cover image. Images are picked from your Media Library and appear in:
To add an image, go to the entity's Edit page. Click Pick from Media Library to browse images you've already uploaded, or visit Generate Images to create AI-generated portraits, illustrations, or scene references.
Worlds are designed to be flexible. Here are five common use cases — each one uses the same underlying entity and relation system, just for different purposes.
Murder Mystery Room
Map suspects, victims, locations, alibis, and the timeline of events. Perfect for collaborative mystery games, escape room design, or crime fiction plotting.
💡 Use Timeline entities with precise timestamps so Sage can check whether alibis are physically possible given travel time between Places.
Family Album & Family Tree
Document your family history — relatives, photos, relationships, migration paths, and oral history. Share it with family members in a private Study Room.
💡 Store each photograph as an Object entity with the photo attached. Add Relations from each person in the photo to the Object entity — Sage can then surface every photo a specific person appears in.
Author Fan World
Build a community lore bible for your favourite author's books. Catalogue characters, places, magical systems, and fan theories alongside canonical facts.
💡 Use Confidence levels to distinguish canonical facts from fan theories. High confidence = directly stated in text. Low confidence = fan reading or speculation.
Fictional Universe / Lore Bible
For authors building their own worlds — map your world's geography, history, factions, characters, and rules into a structured lore bible that Sage can query.
💡 A Worlds lore bible is more powerful than a notes doc because Sage can cross-reference. Ask: 'Which characters were present in Aelindra during the Third Age?' and get a direct answer.
Book Club / Study Room Companion
Build a World alongside your book club's reading — track characters as you meet them, map locations, record timeline events chapter by chapter.
💡 Keep the World private until the club has finished the book, then make it public for future members joining the room.
A Relation is a directed, labelled connection between two entities. Relations are what turn a list of facts into a knowledge graph.
A Relation has:
Navigate to your World, then click the Relations tab. Use the form at the top to create a new relation — pick a From entity, enter the relation label, and pick a To entity. You can also filter the relations list by entity type to focus on just Characters, just Places, etc.
Once your World has entities connected by Relations, it becomes a knowledge graph — a network of nodes (entities) and edges (relations) that can be traversed and queried.
A simplified murder mystery graph
A knowledge graph lets Sage answer multi-hop questions that a flat notes document cannot:
Sage is NotaeLibrary's AI — and Worlds are one of the primary sources it can query. When you ask Sage a question in a Study Room that has a World attached, it has direct access to all your entities and relations.
The richer your World, the better Sage's answers. A few principles:
Structured beats prose
A Character entity with a clear description and linked Relations is more queryable than a long unstructured note about that person. Break information into entities and link them.
Timestamps unlock gap reasoning
Without times, Sage knows order. With times, Sage can compute durations, overlaps, and whether events are physically compatible.
Confidence signals uncertainty
Sage will caveat answers that draw on low-confidence entities, so you know which conclusions are solid and which are speculative.
Add a World Overview Context entity
Create a Context entity titled "World Overview" with a 2–3 sentence summary of what your World is about. This helps Sage frame answers correctly without you having to explain the context every time.
Every World has a visibility setting:
Private
Only you (the owner) and Study Room moderators can view the World and its entities.
Best for: personal research, draft work, sensitive family information, evolving murder mystery theories.
Public
All Study Room members can browse the World's entities and relations via the public lore browser.
Best for: shared book club wikis, fan room lore bibles, completed family trees.
Ready to build your first World? Work through this checklist: