Roadmap
This page lists items on the table for future versions. None of these are promises. Some may ship; some may not; some may change shape entirely. Treat the roadmap as direction, not commitment.
Items are loosely grouped by area and by how concrete they are.
Story bible
Section titled “Story bible”- Tag filter chips on Characters / Locations / Objects. Worldbuilding already has filter chips for tags. The other entity grids don’t yet. Adding them is mechanical work — the pattern’s been established.
Curated project-wide tag vocabulary— shipped (Settings → tag vocabularies; TagEditor canonicalises against it). Deliberately ships EMPTY: a starter default set was considered and rejected — your tags are your own.
Editor and writing
Section titled “Editor and writing”- Sidebar arrow-key navigation. Items in the sidebar are reachable by Tab but arrow keys don’t move within a list. A roving-tabindex pattern would fix this. Accessibility win for keyboard-first users.
- Search snippet highlight offset fix. Some snippets with multi-space runs show highlights off by a character or two. Small fidelity issue, fix planned.
- Token budgets per feature. A per-feature monthly cap that pauses AI calls when exceeded, to prevent surprise bills on cloud providers.
- Embedding rebuild status. The auto-rebuild RAG indicator could be more visible when it’s actively running.
- Finer per-feature engine tuning. Every tunable (temperature, samplers, think) now lives on the engine preset a feature routes to (AI Settings → Routing by feature) — one source, editable today. The remaining roadmap idea is finer grain: distinct preset defaults per feature class (JSON-strict extraction vs prose critique) shipped out of the box, so new users get sensible splits without hand-tuning.
Export
Section titled “Export”- Direct Kindle KFX export. Currently the path is EPUB → Kindle Previewer. A direct export would be a quality-of-life win for indie publishers.
- Print-ready PDF layout knobs. Trim size, margins, page numbering style — currently the PDF export uses sensible defaults. Customisation is on the table.
- Story Bible export. Some writers want to export their character and worldbuilding sheets alongside the manuscript. Format and inclusion logic TBD.
Analysis
Section titled “Analysis”- Configurable style-metric thresholds. Currently the per-chapter style table flags outliers against a fixed baseline. Letting the user define their own “this is acceptable” range would make it more useful.
- Trend lines on the pace chart. A simple regression line showing whether you’re accelerating, holding, or decelerating.
App-wide
Section titled “App-wide”- Multi-window support. Open a Character page in one window and write in another. Currently the app is single-window.
- Cross-machine sync. Right now the path is “point cloud sync at the autosave folder”. A first-class sync built into JustWrite (still local-first, just smarter about merging) is on the wishlist.
- Native dark-mode for charts and graphs. Some chart and graph surfaces use light-mode colours that don’t quite work in dark mode.
- Curated shadow tokens. Several places use ad-hoc shadow rgbas; a
--shadow-soft/--shadow-mediumtoken pair would unlock proper dark-mode shadow behaviour.
Things deliberately not on the roadmap
Section titled “Things deliberately not on the roadmap”A few things JustWrite isn’t planning to add. Worth saying so:
- A cloud account. The local-first model is the whole point. There may be cross-machine sync (see above) but there will not be a JustWrite account you log into.
- A built-in publishing pipeline. JustWrite produces files (EPUB, PDF, DOCX). What you do with them — Amazon, Apple Books, Kobo, IngramSpark — is your business.
- Real-time collaboration. Same reasons as the no-cloud-account decision. A snapshot you hand to a collaborator works; merging concurrent edits across the wire does not.
- Lock-in to any single AI provider. Every AI feature works with whichever provider you choose. That includes local engines that never call out to the internet.
How to suggest something
Section titled “How to suggest something”Open an issue on the JustWrite GitHub repository (link from the main README). Concrete, narrowly-scoped requests get traction faster than broad ones. “Tag filter on Characters” is easier to act on than “make Characters better.”