Migrate your manuscript
Bring a PDF, Microsoft Word manuscript, or Scrivener project into Obsidian’s Markdown + YAML note format — in three staged steps, with your AI of choice doing the heavy lifting.
Stage 1 — Prepare and export your files
Radial Timeline reads plain Markdown notes, so migration starts with clean source files in one predictable place. Whether your book lives in a PDF, a Word file, or a Scrivener project, the goal of this stage is the same: one vault folder, one book folder inside it, and your scenes in a knowable order.
Set up the folders
1.
Create a vault folder — the folder you’ll open in Obsidian. Inside it, create a book folder that holds nothing but this manuscript.
2.
Add your source to the book folder: a single manuscript PDF, or scene/chapter files exported from Scrivener or Word.
3.
Name scene files in numerical order (01, 02, 03…). If your file names don’t carry the order, you must add a TOC note listing the reading order matched to each exact file name.
4.
From Scrivener, export scenes as plain text or Markdown, including synopses and notes. If you created custom meta-data fields, export those too — Stage 3 maps them into YAML so nothing you tracked is lost.
Know your source
From PDF — the lossiest path. Extraction can drop italics and scene breaks, and running headers or page numbers can leak into the text. Keep the PDF as the only file in the book folder; the migration prompt strips headers, footers, and page numbers and keeps emphasis intact.
From Word — chapters usually live as heading styles. Export as .docx; headings become the split points, and italics and scene breaks survive the trip.
From Scrivener — the cleanest path, since scenes already live as separate documents in the Binder. Export with synopses, notes, and your custom meta-data included.
Chapters, parts, and acts — migrate chapter-by-chapter first and split into scenes later; the timeline is structure-agnostic. Radial Timeline defaults to three acts, and Settings can raise the act count to match bigger structures — though a very large number gets impractical. Pick a practical act count, and keep the original division as its own field: the Cyclops scene might carry Act: 2 plus Book: 9.
odyssey-vault
# One vault · one book folder · scenes in order
Odyssey Vault/ ← open this folder as your vault
└─ The Odyssey/ ← one folder per book
├─ 01 Invocation.md
├─ 02 The Council of the Gods.md
├─ 03 Telemachus Sets Sail.md
├─ …
├─ 09 The Cyclops.md
├─ TOC.md ← required if names aren’t numbered
└─ scrivener-fields.csv ← optional custom-field export
✓ every scene has one home — ready for Stage 2
Stage 2 — Convert to Markdown + YAML
One scene, one note. Each note opens with a YAML frontmatter block — placed before any other text — that Radial Timeline reads to draw the ring. The prose below it stays exactly as you wrote it.
Required fields
Class, Act, Status, and Subplot — the minimum for a scene to render on the timeline. Default the subplot to Main Plot and refine later.
Common fields
When (the in-world date), Synopsis, Character, Place, Publish Stage, and Due unlock chronological sorting, hover synopses, and progress colors. Characters and places use wiki links, so each one becomes its own note.
09 The Cyclops.md — Odyssey Vault
---
Class: Scene
Act: 2
When: 1178-03-16
Synopsis: Trapped in the cave, Odysseus blinds Polyphemus and escapes beneath the rams.
Subplot:
- Main Plot
- The Journey Home
Character:
- “[[Odysseus]]”
- “[[Polyphemus]]”
Place:
- “[[Island of the Cyclopes]]”
Status: Complete
Publish Stage: Zero
---
“My name is Nobody,” he said, and poured the dark wine again…
Custom fields are safe. Radial Timeline ignores any field it doesn’t recognize, so your own process fields — POV, Tone, or metadata carried over from Scrivener — can live in the same block without breaking anything.
Stage 3 — Populate advanced fields with AI
You don’t have to fill any of this in by hand. Copy the prompt below into your AI of choice — Claude, ChatGPT, or a coding agent pointed at your vault folder — and it processes the whole book in four reviewable stages, from splitting scenes through the advanced template fields.
Advanced template fields
Beyond the basics, scene notes can carry Duration, Type, Shift, Questions, Reader Emotion, Internal, Words, and POV — the fields the timeline’s deeper modes and AI scene analysis build on.
Scrivener custom fields
The prompt maps every exported custom meta-data field to a same-named YAML field. Then bring them into the timeline: Settings → Scene properties → Advanced properties is Radial Timeline’s advanced YAML manager — activate your custom fields there, give each one an icon, and reveal it in the scene hover. A remapping tool can also point legacy or renamed keys at canonical fields without rewriting your files.
The migration prompt
Copy everything in the window below, paste it into your LLM, and point it at your vault folder. It works stage by stage and waits for your approval between stages.
migration-prompt.txt
ROLE
You are migrating a finished manuscript into an Obsidian vault for the
Radial Timeline plugin. Work in stages. Report after each stage and wait
for my approval before continuing. Never rewrite or “improve” the prose.
SOURCE
- The vault folder contains one book folder with the manuscript:
a single PDF, or exported scene/chapter files (.md, .txt, .docx).
- Numbered file names (01, 02, …) define the reading order.
- If names aren’t numbered, TOC.md maps the reading order to exact
file names. If neither exists, stop and ask me for the order.
- If a Scrivener metadata export exists (synopses, notes, custom
fields), load it now and hold it for Stage 4.
STRUCTURE
- Chapters, not scenes? Treat each chapter as one scene note now;
split at scene breaks (***, blank space) in a later pass.
- Radial Timeline defaults to 3 acts; Settings can raise the
act count to match big structures (the Odyssey’s 24 books).
Pick a practical number, and keep the original division in
its own field: the Cyclops scene gets Act: 2 plus Book: 9.
- From PDF: strip running headers, footers, and page numbers;
keep italics as *emphasis*; never let a page break split a
paragraph.
STAGE 1 — SPLIT INTO SCENES
- Create one Markdown note per scene inside the book folder.
- Name notes “NN Scene Title.md” — zero-padded, narrative order.
- Scene prose goes below the frontmatter, unchanged.
STAGE 2 — REQUIRED YAML
Add this frontmatter block to every scene note:
---
Class: Scene
Act: 1
Status: Complete
Subplot:
- Main Plot
---
Set Act by structural read; if the book runs past 3 acts, raise the act count in Settings to match. Flag uncertain calls.
STAGE 3 — CORE METADATA
- When: the in-world date, YYYY-MM-DD (a bare year is enough)
- Synopsis: 1–2 sentences of what happens in the scene
- Character: wiki links — e.g. “[[Odysseus]]”
- Place: wiki links — e.g. “[[Ithaca]]”
- Create a stub note for every character and place you link.
STAGE 4 — ADVANCED FIELDS
- Publish Stage: Zero | Author | House | Press
- Duration, Words, Due, Pending Edits
- Type, Shift, Questions, Reader Emotion, Internal
- Map Scrivener custom metadata to same-named YAML fields —
Radial Timeline safely ignores fields it doesn’t recognize.
RULES
- YAML frontmatter is always the first thing in the file.
- No commas inside Subplot, Character, or Place names.
- Flag guesses (Act, When) for review — never invent silently.
- Finish each stage across the whole book before starting the next.
