Building the Future of Language Reading
Our strategic plan to make Ember a premium language reading environment where users can open real texts, understand them quickly, hear the language, and build long-term fluency.
Strategic Product Pillars
The roadmap is built around one core loop:
Core Completeness
What Ember needs to feel like a complete product. These are the highest strategic priority.
Broader Format Support
Add TXT, FB2, and basic PDF support. EPUB/FB2/TXT treated as native; PDF as supported with limitations.
TTS for Word and Phrase
Pronunciation playback for selected words and phrases. Core language feature, not just accessibility.
Strong Personal Dictionary Foundation
Lemma-based grouping, stable lemma generation, dictionary-entry model, item states (new/learning/known/ignored).
Better Import and First-Reading Flow
Improve file importing, format clarity, reliability, and first-open experience after import.
Starter Content / Sample Library
Sample books or curated starter texts grouped by language. Enables immediate value and better onboarding.
Premium Reading Experience
What makes Ember feel worth paying for. The transition from functional to exceptional.
Audio-Assisted Reading
TTS for current sentence or paragraph, play-from-here, optional sentence stepping, follow-along highlighting.
Review Tightly Connected to Reading
Light, context-aware review derived from real reading. Easy jump back to source book. Reading-first, not flashcard-first.
Session Continuity
Continue where you left off, subtle reminders about saved or due items, session recap, "resume + optional review" patterns.
Progress Visibility
Pages read, books finished, reviewed items, repeated encounters, knownness growth, comprehension progress per book.
Signature Differentiation
What can make Ember meaningfully distinct from competitors. The features that make Ember memorable.
Re-reading Features
Re-reading with remembered items visible, revisit support, "you used to not know these items" indicators, resurfacing of saved language in context.
Smart Language Triage
Accept to learning, mark known, ignore, archive. Possibly suggest actions based on frequency or repeated exposure.
Context-Rich Review Modes
Review from original sentence, review across multiple encounters, compare similar usages, "seen before in these books".
Advanced Expansion
Valuable later additions that should not compete with higher-priority roadmap items.
Parallel Texts
Strong feature for advanced learners. Adds complexity to UX, content structure, and implementation.
Cloze Deletion
Useful but may push Ember toward generic SRS territory. Only after reading and re-reading foundation is strong.
Additional Review Modes
Typing, multiple choice, matching, flashcard variants. Only if they clearly strengthen the reading loop.
Priority Tiers
Tier A — Must Have
Required for Ember to feel like a strong, complete product.
- TXT support
- FB2 support
- Basic PDF support
- TTS for word
- TTS for phrase/sentence
- Strong lemma infrastructure
- Personal dictionary foundation
- Better import flow
- Starter/sample content
- Polished reading-save-lookup loop
Tier B — Should Have
Make Ember feel premium and coherent.
- Audio-assisted reading
- Session continuity
- Reading-connected review
- Progress visibility
- Smart triage for saved items
Tier C — Differentiation
Can make Ember truly memorable.
- Re-reading features
- Context-rich resurfacing
- Stronger accumulation model
- Knownness / repeated exposure logic
Tier D — Later
Advanced expansion when foundation is solid.
- Parallel texts
- Cloze deletion
- Extra review modes
What We're Not Building
These should not be priorities right now:
- Many review modes just for feature count
- Heavy gamification
- Turning Ember into an Anki-like system
- Trying to make PDF equal to EPUB
- Building a huge built-in library before starter content proves value