Changelog
What's in each release of SceneWeaver, newest first — from the earliest build through to the current release.
Version 2.0.11 released 31 August 2026
This release fixes issue D7MC0MCJF52S. In some documents, two built-in tokens placed next to each other with only a space between them — a title page with the copyright year and author next to each other, for example — could cause SceneWeaver to lose track of the correct editing position in Show Text mode. Text entered or edited after those tokens could then be inserted, removed or changed in the wrong place in the underlying document, potentially resulting in corrupted text. Fixed by aligning how SceneWeaver counts adjacent tokens on both the visible and the stored side of the document, so subsequent edits always land in the right position. The same fix has been back-ported to Version 1.
Version 1.0.1X released 31 August 2026
Back-port of the Version 2.0.11 fix for issue D7MC0MCJF52S. Two built-in tokens placed next to each other with only a space between them could cause SceneWeaver to lose track of the correct editing position in Show Text mode, potentially corrupting text entered after them. The Version 1 line only carries the token-counting fix, which is the part of the issue that applies here.
Version 2.0.01 released 25 August 2026
This release fixes issue SRR94MDSA209 corkboard reogranisation. When the manuscript was selected the corkboard ignored drag and drop due to the special chapters having an enforced position. This behaviour was incorrect. This release introduces a blue bar to show exactly where a chapter will end up when it is dragged to a new position. Only normal chapters can be rearranged in this way.
Version 2.0.00 released 21 August 2026
Version 2 is the first paid release of SceneWeaver and the biggest single expansion of what the app can do since it first shipped. Every version-1 feature is still here; version 2 adds a full research workspace, span-anchored notes, book-design pagination for the finished book (not just the manuscript), token-driven prose that follows your entity renames, coexisting spell and grammar highlighting via a one-touch LanguageTool installer, and themes. Version 1 users can stay on Version 1 forever and never pay; Version 2 introduces optional Student and Pro tiers with a one-off payment, and the Free tier continues indefinitely with a three-manuscript cap.
Research Library. Feature. A dedicated area inside your manuscript file for reference material: folders of notes, images, saved links (with their reader-mode text captured for offline reference), quotations, and imported PDF or RTF documents. Every item can be bound to a specific scene, chapter, or entity. Benefit. What you needed while writing that scene is one click away next time you're in it, and none of it accidentally exports with the finished manuscript.
Book design and pagination. Feature. Special-chapter types (front cover, title, dedication, acknowledgements, epigraph, about-the-author, back cover) that mirror how real books are laid out, each with automatic pagination rules on export. Background watermarks per chapter, scene, or section — whichever is the lowest level used — alongside built-in tokens for manuscript metadata (title, author, ISBN, publisher) and front-and-back-matter fields (dedication, acknowledgements, epigraph, about-the-author). Benefit. The exported document reads and looks like a finished book — not a Word document with paragraphs in it — without you having to build the layout by hand in a separate tool.
Convert-to-Token, and the Built-in Tokens panel. Feature. Right-click any character or location mention in your prose and turn it into a merge token that follows entity renames. The Built-in Tokens panel on the right side of the writer window lists every non-entity token available (metadata, front & back matter, copyright, contact) — double-click any entry to insert it at the cursor. Benefit. Rename an entity once and every previously-typed mention updates with it; set your manuscript metadata once in Preferences and every export renders it consistently.
Inline annotations with jump-to-source. Feature. Anchor a note to a specific span of text — a passage that needs a rewrite, a query for a beta reader, a reminder to check a fact. The Notes panel on the right lists them all with jump-to-source, and annotations survive edits to the surrounding prose. Benefit. Revision notes stop cluttering the prose itself, and the annotations layer never accidentally exports.
Grammar highlighting that coexists with spelling. Feature. Optional grammar checking via LanguageTool, installed on demand from inside the app in one click (no external Java hunt if none is present on your machine). Green grammar underlines appear alongside the red spelling underlines without either erasing the other. Benefit. You can see spelling and grammar issues at the same time on the same word, and grammar checking doesn't require any hand-configuration before you can use it.
Themes. Feature. Light, dark, and mid-tone SceneWeaver palettes. Benefit. Whichever writing environment fits the room you're in.
Inline images throughout, exported inline everywhere they can be. Feature. Drop an image into a scene body and it becomes part of the prose, exported inline through DOCX, EPUB, PDF, HTML, Markdown, and LaTeX. Attach images to a character record for a face and a wardrobe, to a location record for a floor plan, or as chapter background watermarks and cover artwork. Benefit. Every image travels inside the manuscript file — no external folder dependencies to break — and rides through every export format that supports images.
Version 2 is available now. See Pricing for tiers — the £19 launch discount on Pro runs until 30 September 2026. Version 1 continues to be free for everyone, no cap, indefinitely — available on the older versions page.
Version 2 development history
The internal build cycles that ran between v1.0.0X (the final v1 release) and v2.0.00 (the launch). None of these were ever the promoted public download — they are the dev history that led to v2.
Version 1.12.042Y citation-processor regression fixes + test-suite cleanup
Final polish on the citations pipeline — regression fixes surfaced by the new tests, and a cleanup pass across the suite — sealing the citations work before the v2.0.00 tag.
Version 1.12.041Y entity dialogs auto-save + Concept type + ignored-restore + sample-from-current
Four related quality-of-life improvements. Entity dialogs now auto-save so you don't lose edits by clicking away. A new Concept entity type joins Characters, Locations, Items, and Organisations, for the abstract nouns a story needs to track. Ignored manuscripts can be restored to the drop-in list. And a new sample manuscript can be spun up from your current one.
Version 1.12.039Y citations overhaul (CSL) + index page-estimator + sweep of 016Y–038Y
The biggest single content pass of the v2-dev cycle. Full Citation Style Language support arrived, replacing an earlier ad-hoc citation system with a proper CSL 1.0 engine and ten bundled style files. The Index gained a page-count estimator so it can render meaningfully before real pagination arrives in v3. Also rolls up twenty-two internal iterations (016Y through 038Y) that never received their own external version bump.
Version 1.12.015Y shipped sample covers cleared + samples now tracked in git
The shipped sample manuscripts had leftover cover images from earlier internal use; those are cleared. Samples themselves are now tracked in git so a fresh install picks up the correct baseline.
Version 1.12.014Y F7 tokens panel expansion + cover watermarks 100% + author-notes tick fix
The F7 Tokens panel got a proper expansion of what it can insert. Cover-page background watermarks now render at 100% opacity as the design intended. And an incorrect tick state on Author's Notes was fixed.
Version 1.12.013Y word-count perf + Timeline auto-refresh on add-node
Word-count performance improved for very large manuscripts. The Timeline now refreshes automatically when a new node is added, so you don't need to switch away and back to see it appear.
Version 1.12.012Y word count resolves tokens before counting
Word counts now resolve merge tokens to their expanded values before counting, so a manuscript that leans heavily on entity tokens gets an accurate total rather than one that under-counts every token as a single word.
Version 1.12.011Y Timeline shows Unplaced Content
The Timeline view gained an Unplaced Content label plus a prose-specials filter, so content that hasn't been placed on the timeline is visible instead of quietly invisible.
Version 1.12.10AY list-marker font fix + Bullet/Number/Align toolbar pressed-state
List markers now use the same font as the list content. Bullet, Number, and alignment toolbar buttons now show a proper pressed state so the current formatting is visible at a glance.
Version 1.12.0ZY trailing-empty-block font matches manuscript default
A trailing empty paragraph at the end of a scene no longer keeps the font of whatever formatting run preceded it; it reverts to the manuscript default so the next paragraph you type looks right.
Version 1.12.0YY editable font-size spinner + size-based heading detection
The font-size control became an editable spinner rather than a fixed dropdown. Heading detection now works from typed font size as well as explicit style choice.
Version 1.12.0XY Front/Back Cover auto-pin Image Library + no inline-insert
Opening a Front or Back Cover chapter auto-pins the Image Library panel so cover artwork is one click away. Inline image insertion is disabled on cover pages — covers are background-image-only by design, and the inline-insert affordance was misleading.
Version 1.12.0WY read-only pages get Remove-background right-click menu
Read-only pages (cover chapters, section headers) now expose a right-click Remove-background option so a wrongly-applied background can be cleared without having to work around the read-only state.
Version 1.12.0VY Apply-as-Background inline menu + Cover-page tint + runbook fix
Images in the Image Library gained an Apply-as-Background inline menu action. Cover pages picked up a tint control for background artwork. UAT runbook wording corrected on a related step.
Version 1.12.0TY v3-dev candidates + colour-blind audit logged
No user-visible change. Logged v3-dev candidates for a custom title bar and a full colour-blind audit as future work items rather than blocking the launch.
Version 1.12.0SY Windows title bar honours Dark + High Contrast themes
Windows' native title bar now honours the SceneWeaver Dark and High Contrast themes instead of always painting a light title bar over the themed window body.
Version 1.12.0RY Research Library panel honours Dark + High Contrast themes
The Research Library panel had been hard-wired to light-theme colours; it now honours the active Dark or High Contrast theme like every other panel.
Version 1.12.0QY install UUID generated eagerly on first launch
The per-install UUID (used for anonymous telemetry when opted in, and for licence-install binding on paid tiers) is now generated on very first launch rather than lazily on first use.
Version 1.12.0PY default Windows projects folder now lives in Documents
The default projects folder on Windows moved to Documents\SceneWeaver, matching where a Windows user expects a writing app to save their work.
Version 1.12.0NY UAT runbook Windows user-state path corrected
Internal UAT runbook fix — the documented Windows user-state path was wrong, causing testers to look in the wrong place.
Version 1.12.0MY Windows build script survives empty pip cache
Build-tooling fix — the Windows build script now handles an empty pip cache instead of failing part-way through.
Version 1.12.0KY menu bar + menus now honour Dark / High Contrast palettes
The menu bar and pop-up menus previously kept their platform-default light appearance regardless of the SceneWeaver theme. They now honour Dark and High Contrast alongside every other UI surface.
Version 1.12.0JY block-style combo no longer misreads plain paragraphs as Author's Note
The block-style combo box was falsely reporting some plain paragraphs as Author's Notes. It now reads the paragraph's actual style, not a heuristic that could be wrong.
Version 1.12.0HY node loads honour the persisted Show Text preference
Opening a node now uses whichever Show Text mode you last chose (raw tokens vs resolved text) rather than reverting to the default every time.
Version 1.12.0GY plain-English consent copy
The first-launch telemetry consent dialog copy rewritten from technical wording ("telemetry", "diagnostic events") into "usage information", with a clearer local-only clarity note. What the app does hasn't changed; how the app explains it has.
Version 1.12.0FY Smart Collections refresh after structural + status changes
Smart Collections now refresh automatically after a structural change (adding, moving, or deleting scenes) or a status change, so the collection matches the manuscript without a manual refresh.
Version 1.12.0EY §6 library binding runbook wording + Future Feature 64
Internal-only. Sixth-section library-binding runbook wording clarified, and a new future-feature slot (64) reserved.
Version 1.12.0DY Manuscript Settings dialog gains a Title field
The Manuscript Settings dialog now has a proper Title field, exposing the setting that had previously only been reachable via the underlying data model.
Version 1.12.0CY entity rename rewrites existing tokens
Renaming an entity now updates every existing merge token that references it. Previously the tokens kept the old name until each was re-typed.
Version 1.12.0BY inline images survive promote/demote in Show Text mode
Inline images placed in the prose now survive a chapter promote or demote operation while in Show Text mode, instead of being silently lost.
Version 1.12.0AY image selection overlay follows alignment changes
The selection overlay around a selected image now moves with the image when its alignment changes, instead of staying at the pre-change position until a re-select.
Version 1.12.09Y file-system image drop inserts the image, not the path
Dragging an image from the file system into the prose editor now inserts the image itself, not the file-system path as plain text.
Version 1.12.08Y numbered-list Continue / Reset no longer stacks indent
The Continue and Reset actions on numbered lists no longer add cumulative indentation with each use.
Version 1.12.07Y UAT runbook rewrites from UAT-2 notes
Internal-only. UAT runbook rewritten to fold in the corrections that surfaced during the second UAT round.
Version 1.12.00AE block-style combo definitive; annotation-card crash fix
The block-style combo is now the definitive source of truth for the current paragraph style. A crash on the annotation card display was also fixed.
Version 1.12.00AD block-style combo now shows the chosen heading and quote
The block-style combo now reflects the paragraph's Heading level (H1–H4) or Quote style rather than always saying "Body".
Version 1.12.00AC Move Epigraph between Front and Back matter
Epigraph chapters can now be moved between Front matter and Back matter without having to delete and recreate them.
Version 1.12.00AB About the Author special chapter
Added the About the Author special chapter type, joining Dedication, Acknowledgements, Epigraph, and the cover chapters in the front-and-back-matter set.
Version 1.12.00AA new-manuscript starts empty; Front/Back Cover read-only
A new manuscript now starts as an empty tree rather than pre-populated with a template. Front and Back Cover chapters became read-only in the tree (their content is set via the cover-page controls, not free-form text editing).
Version 1.12.00Z telemetry consent dialog now offers both opt-ins pre-ticked
First v2-dev version. The first-launch telemetry consent dialog now presents both the bug-report and anonymous-usage opt-ins with the checkboxes pre-ticked, so agreeing is one confirmation rather than three — while every user's ability to say no on either or both remains exactly as before.
Version 1.0.0Y 6 August 2026 — internal polish, not released
Two small polish fixes bundled onto the pre-launch live line. Never promoted to the public download — v1.0.0X remained the promoted final v1 release, and the fixes here were carried into v2.0.00 directly.
Corkboard drill-in restored on special chapters. Feature. Double-click on a Prologue, Epilogue, front cover or back cover card in the Corkboard now drills into that chapter’s single-focus view, the same as any regular chapter card. Benefit. Every childless card in the Corkboard is reachable by double-click again — previously the special-chapter cards were dead to the mouse.
About dialog copy honest about LanguageTool setup. Feature. The About dialog no longer says LanguageTool must be installed separately — it now names the one-touch “Set up LanguageTool” button in Preferences that downloads and configures it for you. Benefit. The copy matches what the app actually does; new users aren’t sent looking for an install step that isn’t there.
Version 1.0.0X released 4 August 2026
The final Version 1 release before the Version 2 launch. Two writing-workflow improvements: a first-launch feedback question that makes anonymous usage sharing a proper conversation instead of a hidden setting, and a much smarter treatment of entities that only live in part of your manuscript.
Story-scope entity awareness. Feature. Characters, locations, items and organisations can each carry an Introduction Node and an Exit Node — the point at which they enter and leave the story. Every entity-aware feature now honours those boundaries: the Data Quality report's Conflicting Keywords tab, the First-mention and Post-exit continuity checks, the Entity Usage report's Used / Unused status and its per-node occurrence list, the token resolver expanding <<character:name>> in Show Text mode, and the Locations / Characters / Items / Organisations side panels' used indicator. Benefit. Two entities that share a keyword but live in different parts of the story — a Colonel who dies in Act 1 and a Colonel who arrives in Act 3, or two archives that both call themselves "the Archive" — are now attributed correctly wherever they appear. A keyword appearing in a scene is credited to the entity that's actually present, not the one that sorts alphabetically first.
First-launch feedback question. Feature. On very first launch of SceneWeaver (and again when a future major version arrives), a small dialog explains what SceneWeaver can send anonymously about how the app is running, what it never sends, why the team asks, and why sharing is worth it. Enable or Decline — either choice is remembered, and can be changed any time in Settings → Privacy & Telemetry. Benefit. The feedback option no longer lives in a settings screen a first-time user might never open. Your consent is a proper conversation with a plain-English promise, not a buried checkbox.
AppImage runs cleanly on modern Linux distributions. Feature. The Linux .AppImage now uses the current-generation AppImage runtime. Benefit. Ubuntu 24.04+, Debian 13+ and other current distros run SceneWeaver out of the box without needing to side-install the legacy libfuse2 package. Older distros still work as before.
Free for everyone, as always. Available on the older versions page.
Version 1.0.0W entity usage report -- per-node token resolution
Fourth pass through the Ashcombe manuscript's shared-keyword bug. The Entity Usage report was resolving merge tokens without knowing which scene a token appeared in, so <<archive:name>> in Scene 1 was expanding to whichever entity happened to sort first — polluting the resolved text before the range check even ran. The report now resolves tokens per-leaf using the leaf's own node id, so downstream attribution operates on truthful text. Superseded within the same day by v1.0.0X's side-panel finish.
Version 1.0.0V entity report + drill-down node-range aware
Third pass through the same bug family. The Entity Usage report's Used flag and the per-node Occurrences drill-down still didn't respect the intro / exit ranges after v1.0.0T. Introduced a shared attribution helper on the entity service so the same "does this entity own this match at this node" rule the continuity checks use is now applied to the report too. Version letter U skipped by convention (Crockford base 32 elides U to avoid glyph confusion with V).
Version 1.0.0T entity resolver + continuity + side panel
Second pass through the entity node-range fix, after UAT of v1.0.0S surfaced four more places that shared the same blind spot. The merge-token resolver, the First-mention Order continuity check, the Post-exit Appearances continuity check, and the entity side panel's used indicator all learned to attribute a matched keyword to the entity that's actually active at the current node — not the one that happens to sort alphabetically first.
Version 1.0.0S entity highlighter + conflicts report
First pass on entity node-range awareness. The prose-editor highlighter and the Conflicting Keywords report both learned to consult each entity's Introduction Node and Exit Node before attributing a keyword match. UAT on a real manuscript surfaced further keyword-matching call sites that needed the same treatment; the remaining fixes ship in v1.0.0T through v1.0.0X.
Version 1.0.0R first-launch feedback & AppImage runtime
A first-launch feedback question and a Linux install-path fix.
First-launch anonymous-feedback consent. On very first launch (and again on any future major version bump), a plain-language dialog explains what SceneWeaver can send anonymously, what it never sends, and why sharing helps — with an explicit Enable / Decline. The setting remains editable in Settings → Privacy & Telemetry. Replaces the previous flow where the option was a hidden checkbox nobody was ever likely to find on their first day.
AppImage rebuilt against the current-generation runtime. Ubuntu 24.04+, Debian 13+ and other modern distributions run the Linux AppImage without needing the transitional libfuse2 package. Older distributions still work as before.
Version 1.0.0Q released 19 July 2026
Previous stable release, rolling up nineteen patches since v1.0.05.
Structural + semantic keyboard navigation. Move around your manuscript at the level you actually think in. Scene / section navigation via Ctrl+PgUp / Ctrl+PgDn walks every level of the tree (not just scenes); heading navigation via Alt+Up / Alt+Down crosses scenes if the current one has no more headings; entity mention navigation via Ctrl+Alt+Left / Ctrl+Alt+Right jumps between references to a character / location / item / organisation; spelling and grammar issue navigation via F3 / Shift+F3 steps between underlined issues in the current scene.
Prose-level keyboard shortcuts. Sentence navigation via Alt+Left / Alt+Right; paragraph navigation via Ctrl+Up / Ctrl+Down; every Shift+ variant to extend the selection. Fixed a Windows-only issue where modifier keys carried extra platform flags that hid Ctrl from the shortcut layer.
Bullet and numbered list buttons rebuilt. Toggling Bullet or Numbered List on a whole selection now turns the whole selection into a list; toggling again turns it off. Fixes a case where multi-paragraph selections were silently replaced with a single empty list stub. Auto-continue numbering means adding a paragraph back into a numbered list resumes the sequence instead of resetting to 1. Right-click a list item for manual Continue numbering or Reset to 1 options. Layout no longer leaves extra indent after list changes.
Reports scan resolved prose, not raw merge-token markup. Overused-word and phrase reports used to count words that only appeared inside merge tokens like <<character:name>>. Every report now runs against the resolved text — the same words your readers see.
Prose-editor stability. A rare edge case that could clear paragraphs when switching between the resolved and raw token views has been closed out across multiple patch releases. Editor undo, cursor placement, and format state now survive round-trips through the resolved / raw token views cleanly.
Version 1.0.0P prose-wipe narrowing
Further narrowing of a rare prose-wipe that could clear paragraphs when switching nodes, based on UAT reproduction from the previous patches.
Version 1.0.0O prose-sync defence softened
Softened the prose-sync defence so a token-resolve failure falls back to the original text rather than risking a wipe. Companion diagnostic in v1.0.0N.
Version 1.0.0N prose-sync diagnostic log
Diagnostic log lines added around the prose-sync path so any remaining wipe corner leaves a forensic trail. No visible change; groundwork for the v1.0.0O + v1.0.0P fixes.
Version 1.0.0M F3 rescoped to current scene
Rescoped F3 / Shift+F3 issue navigation to the current scene only. Spelling and grammar highlighters only run against the scene loaded in the editor, so cross-scene walking was scanning scenes whose highlighters had never fired — wasted work, guaranteed to find nothing.
Version 1.0.0K F3 walks on-screen underlines
F3 now walks the actual on-screen underlines instead of a stale scan cache.
Version 1.0.0J F3 spelling & grammar navigation
Spelling and grammar issue navigation via F3 / Shift+F3. Step to the next or previous underlined issue with the keyboard, no mouse required.
Version 1.0.0H entity mention navigation
Entity mention navigation via Ctrl+Alt+Left / Ctrl+Alt+Right. Jump between references to a character, location, item, or organisation across the whole manuscript.
Version 1.0.0G prose-wipe root cause
Root-cause fix for a prose-wipe that could clear the scene body when switching between the resolved and raw token views. Companion to the defensive guards in v1.0.0E and v1.0.0F.
Version 1.0.0F prose-wipe guard extended
Extended the prose-wipe guard to scenes containing no merge tokens — the original guard only covered token-bearing scenes.
Version 1.0.0E prose-wipe defensive guard
Defensive guard added around the prose-wipe case while the root cause was still under investigation. Superseded by the real fix in v1.0.0G.
Version 1.0.0D list-toggle indent fix
Bullet and numbered list changes no longer leave extra indent that only cleared after navigating away and back. Visual layout now updates immediately.
Version 1.0.0C auto-continue numbering & reset
Auto-continue numbering: adding a paragraph back into a numbered list resumes the sequence instead of resetting to 1. Right-click a list item for manual Continue numbering and Reset to 1 options.
Version 1.0.0B list-toggle rewrite
Bullet and Numbered List buttons rebuilt as paragraph-level toggles: click on a whole selection to turn it into a list, click again to turn it off. Fixes a case where multi-paragraph selections were silently replaced with a single empty list stub — losing the selected content.
Version 1.0.0A heading navigation
Heading navigation via Alt+Up / Alt+Down. Walks across scenes if the current one has no more headings in the requested direction.
Version 1.0.09 whole-tree structural navigation
Ctrl+PgUp / Ctrl+PgDn extended to walk every level of the manuscript tree — sections within a scene, scenes within a chapter, chapters within the manuscript — not just scenes.
Version 1.0.08 scene navigation
Scene navigation via Ctrl+PgUp / Ctrl+PgDn. Superseded the same day by v1.0.09's whole-tree walk.
Version 1.0.07 Windows modifier-key fix
Windows fix for the v1.0.06 sentence / paragraph shortcuts. On Windows the arrow-key events sometimes carry extra platform flags (KeypadModifier, native scancode bits) that hid the Ctrl modifier from the shortcut layer.
Version 1.0.06 sentence & paragraph navigation
Prose keyboard navigation: sentence via Alt+Left / Alt+Right, paragraph via Ctrl+Up / Ctrl+Down, plus every Shift+ variant to extend the selection.
Version 1.0.05 released 29 June 2026
The first maintenance update on top of v1.0. Polishes export layout, fixes a drag-and-drop quirk in the manuscript tree, and tightens startup.
Manuscript-tree drag and drop now behaves intuitively in every drop position. Drop a scene onto another scene to insert it above; drop into the gap between two chapters to stay at the end of the current chapter; drop just below a chapter row to land at the top of that chapter's scene list. Previously some drop positions silently did nothing, or pulled the scene into the wrong chapter.
Heading alignment carries through to every export format. Centred, right-aligned, and justified headings (Heading 1-6) now keep their alignment in DOCX, EPUB, PDF, and HTML output — so a centred title page renders correctly in the final document. PDF exports also stop forcing each title-page line onto its own page, and HTML exports preserve paragraph first-line indents.
Clean shutdown if the main window fails to load. A rare startup failure (typically caused by a corrupted manuscript file or a missing optional dependency) now reports the error and exits cleanly, instead of leaving a zombie SceneWeaver process running in the background.
Free for everyone, as always. Download v1.0.05 for Linux or Windows.
Version 1.0.04 tree drag-drop resolvers
Manuscript-tree drag-and-drop position resolvers rewritten so scenes land where you expect them to. Superseded the same day by v1.0.05's edge-case fix for the "drop just below a chapter row" case.
Version 1.0.03 export layout & startup robustness
Export layout polish: heading alignment (centred, right-aligned, justified for Heading 1-6) now carries through DOCX, EPUB, PDF, and HTML output so centred title pages render correctly in the final document. PDF exports stop forcing each title-page line onto its own page; HTML preserves paragraph first-line indents. Startup robustness: a rare MainWindow-load failure now reports the error and exits cleanly instead of leaving a zombie process behind.
Version 1.0.02 paragraph style & OS-aware fonts
Backport wave from ongoing v2-dev work. Paragraph style dropdown rewritten to override every property (margins, text-indent, alignment) so a leftover setting from the previous style no longer bleeds through. Title block improvements. OS-aware default body font — Arial on Windows, Helvetica on macOS, Liberation Sans on Linux — so a Windows-only font like Segoe UI no longer resolves to whatever unrelated font Qt happens to have on hand. Citation chip stability + a batch of dialog fixes.
Version 1.0.01 test-suite housekeeping
Test-suite housekeeping after a Licences page heading rename. Captured KB-001 (telemetry buffer not cleared after a successful flush) as a known follow-up; no visible change for writers.
Version 1.0 released 10 June 2026
The first public release. Version 1 ships the complete SceneWeaver feature set: a structured manuscript tree for chapters, scenes, and sections with automatic word-count roll-up; rich entity records (backstory, description, and usage tracking) for characters, items, locations, and organisations; the full continuity-validation suite; every report (Statistics & Readability, Sentence Breakdown, Repeated Phrases, Entity Usage, Data Quality & Completeness, and the Writing Coach dashboard); eight export formats (DOCX, PDF, EPUB, HTML, Markdown, plain text, Fountain, LaTeX) with Compile Presets; citations and bibliography with live Zotero import; the timeline view; Restore Points and the Snapshot Comparison Viewer for content history; Revision Rounds for editorial-pass status tracking; the Manuscript Health diagnostic; and forward-only data compatibility with forced pre-migration backups.
Inline spell-check ships with UK and US English Hunspell dictionaries bundled in. Optional inline grammar checking is provided by LanguageTool — installed on demand from inside the app with one click, no Java setup required if none is present on your machine.
Free for everyone. No accounts, no activation, no watermarks. Manuscript content never leaves your machine; bug reports and anonymous usage telemetry are opt-in and off by default. Download for Linux or Windows now. A macOS build is demand-led — see the Download page if you'd run it on a Mac. For the full feature breakdown, see Features.
Version 0.16 pre-release integrity hardening
The final layer of trust before version 1. Reference-integrity audits across every node reference; round-trip and export golden-file tests pinning DOCX and Markdown semantics; schema-version stamping with forced backup before any migration; a Project Health diagnostic that surfaces continuity, reference, and backup facts at a glance. The layered app-icon system and the public marketing site landed here too.
Version 0.15 licensing, performance, integrations
LanguageTool grammar decoupled into an optional external integration. Release licence compliance — on-disk licence file + Help → About → View Licences. Per-rule continuity validation with a new dangling-references check. Hierarchical report scope (whole manuscript / chapter / scene / section). Runtime performance profiling with a diagnostics dialog. Status-bar indicators for LanguageTool and Zotero. Bundled sample-manuscript library reworked.
Version 0.14 citations, scoped output, timeline polish
Bibliography and citation management with live Zotero import; four citation styles (Chicago, APA, MLA, Harvard); Insert and Edit Citation dialogs; multiple-source and secondary "cited in" tokens. Timeline drag-and-drop reordering. Version labels and milestones. Publisher target presets. Scoped and review printing.
Version 0.13 reference consistency & round-trip archives
Node reference consistency across every structural mutation. Enhanced beta-reader / DOCX round-trip archive. Markdown round-trip archive. Split-content formatting preservation fix.
Version 0.12 revision workflow & timeline
Full revision workflow. Enhanced continuity validation and entity signatures. DOCX round-trip import. Timeline / chronology view.
Version 0.11 Fountain, continuity, unified dialogs
Fountain (screenwriting) export. Continuity validation. Per-node status labels. Unified export and import dialogs.
Version 0.10 comments, markdown import, consistency
Inline comments and annotations. Markdown import. Style and consistency checks. Platform hardening pass.
Version 0.9 EPUB, corkboard, beta-reader
EPUB 3 export. Corkboard / card view. Smart collections. Beta-reader package export.
Version 0.8 targets, coach, more exports
Additional export formats (HTML, plain text, and related). Writing targets and progress tracking. Writing Coach dashboard. Focus / composition mode.
Version 0.7 PDF, presets, restore points
PDF export with print preview. Compile presets. Restore Point productisation. Snapshot Comparison Viewer. Version and build visibility. Expanded report filters and drill-down.
Version 0.6 data quality & merge tokens
Data Quality and Completeness reporting. Merge-token disambiguation and indexed tokens. Panel and highlight UX hardening. Report export hooks.
Version 0.5 reports & search
Reports menu and access. Statistics & Readability. Entity Usage. Sentence Breakdown. Repeated Phrases. Editor search and search-and-replace.
Version 0.4 entities
Characters. Items. Locations. Organisations and factions. Shared entity store, usage panel, keyword matching, filtering, entity detail view, merge-token support.
Version 0.3 settings, persistence, samples
Window size and geometry persistence. Smart manuscript auto-open on startup. Settings and preferences windows. Dual sample manuscripts and the dedicated Help dialog.
Version 0.2 reordering & DOCX export
Structural reordering. DOCX export delivery. Test execution and regression coverage. Requirements consolidation pass.
Version 0.1 core foundation
Manuscript projects: create and open, persisted in SQLite. Hierarchical writing nodes — chapter, scene, section — with add, edit, split, and remove. Desktop tree plus editor layout. Automatic word-count display.