Identity & visibility
- id — stable identifier (UUID-like).
- title — human-readable title (can be empty).
- slug — optional URL slug, used for public / unlisted links.
- scope —
private/unlisted/public.
Docs · Core— Fragments as small units of living knowledge
Fragment basics
A Fragment is the basic page unit in Fragment.place — one note, one URL, one small piece of living knowledge. Most of the time, you just write the main narrative in Markdown. When you need more structure, you add tiny YAML config blocks and Scene / Flow diagrams alongside the text.
「Fragment」は、Fragment.place における1 ページ分の最小単位です。 日記、議事録、仕様書、レビュー、講義ノートなど、 ほとんどのコンテンツはこの 1 単位で扱われます。そのうえで必要なときだけ、 YAML や図（Scene / Flow）をそっと足していくイメージです。
Journals, specs, lecture notes, reviews, mini-essays — they all share the same Fragment model: header + content + a few small blocks.
A Fragment in three layers
Internally, a Fragment is simple. You can think of it as three layers:
[Header] : id, title, scope, tags, timestamps [Content] : main Markdown body you read and write [Blocks] : small YAML / Prism / Scene / Flow chunks → One Fragment = one page + a few structured islands of config and diagrams.
You don't need to remember the full schema while writing. It is enough to think: title + body + a few blocks nearby. The rest of this page makes that mental model more precise for designers, developers, and future integrations.
Conceptual structure (spec-level view)
Here is a high-level, schema-like view of how a Fragment can be represented. Actual storage details can evolve, but this gives you a stable reference:
fragment:
id: "def5d6b2-d52c-4b96-ab6d-52808f051d11"
title: "Example Fragment"
slug: "example-fragment"
scope: "private" # private | unlisted | public
tags: ["journal", "study"]
author: "@you"
created_at: "2025-01-01T10:00:00Z"
updated_at: "2025-01-03T08:12:00Z"
content:
markdown: |
# Heading
This is your main note, written in Markdown.
You can insert `yaml`, `scene`, and `flow` blocks.
blocks:
- kind: "yaml"
label: "config"
value: |
config:
scope: "unlisted"
default_prism: "study"
- kind: "prism"
label: "study-partner"
value: |
prism:
persona:
name: "Study partner"
tone: "kind, patient"
- kind: "scene"
label: "layout-diagram"
value: |
scene:
# ...
meta:
versions:
count: 5
stats:
views: 42
reactions: 3In Studio, you mostly interact with Markdown and fenced blocks. The header, meta, and block list are what let the reader, Explore, and AI tools make sense of that same page.
Core header fields of a Fragment
These are the fields you're most likely to care about when designing integrations, readers, or exports. Names may evolve, but this is the core shape.
Classification
- tags — free-form tags for grouping and search (3–8 per Fragment is a healthy range).
- author — handle or user identifier.
- collections (planned) — membership in booklets / zines / sets.
Timestamps & versions
- created_at / updated_at — when this Fragment was created and last saved.
- versions — version count and pointers for history / rollback.
- stats — lightweight stats such as views or reactions.
AI & layout hints (via YAML)
- prism — default Prism profile name to prefer when calling AI on this page.
- layout — hints like
article,slides, ormixedfor readers. - links (planned) — cross-page relations and backlinks.
Content and structured blocks inside a Fragment
Inside the page, Fragment keeps the Markdown body and an ordered list of small structured blocks. These blocks are what other features and AI can build on.
Markdown body
This is where you write normally: notes, specs, journal entries, lecture summaries, reviews, and so on.
Fragment treats Markdown as the primary narrative for humans to read. Structured blocks sit near it as small, explicit hints for how to use that page.
Blocks that exist today
block:prism— AI persona, tone, boundaries, and explanation style.block:config— small local config (audience, language, notes).- Scene / Flow — diagrams as text for layouts and processes.
block:tasks,block:calendar,block:relations— emerging family for tasks, schedules, and links.
# Weekly review — work & family ```yaml block:prism persona: name: "Gentle weekly coach" role: "helps me reflect without guilt" tone: style: "calm, concrete" directness: "low" ``` This week I worked on... ```yaml block:scene scene: # Simple layout of "Work", "Family", "Energy" boxes ``` Notes, thoughts, and prompts continue here...
Scopes — private, unlisted, public
Scope controls who can see a Fragment. It is independent from version history or Prism settings.
private
Only you can view the Fragment. Most notes start here and never leave this scope.
unlisted
Anyone with the URL can view it, but it won't appear in Explore. Good for work-in-progress or trusted collaborators.
public
Visible in Explore and other public surfaces. Ideal for essays, templates, and teaching materials.
Where Fragments appear in Fragment.place
Studio
The place where you write and grow a Fragment. Markdown, Prism, and Scene / Flow blocks live in the same editor.
Open Studio →Reader
A calm read-only view for public and unlisted Fragments. Designed for comfortable reading, sharing, and printing.
Browse public Fragments →Deck / Export
Features that turn a Fragment into slides or export it as PDF / Markdown — all built on top of the same page model.
Learn about exports →How a Fragment travels together with AI
When you ask AI to help with a Fragment, Fragment.place doesn't just send a raw prompt. It can package the body, Prism, diagrams, and config together as one unit.
[Body] : The main Markdown note. [Prism] : How AI should behave here (persona, tone, boundaries). [Config] : Local hints (audience, language, layout). [Scene/Flow] : Diagrams for structure and processes. → When a Fragment is passed to a model, these pieces can be combined, so AI reads the same context you do — not just an isolated paragraph.
This is the core idea of Fragment: one page, one shared structure for you and your AI tools. Other features (Explore, templates, decks) are layered on top of this same unit.
Create your first Fragment
You don't need the perfect system on day one. Start with a short note, leave the title or tags blank if you want, and add YAML, Prism, or Scene / Flow only when they become useful. The Fragment model is designed to grow as your thinking does.