Capture From Safari
Save selected Japanese directly from a Safari context menu while retaining the page, surrounding sentence, paragraph, source URL, and time of capture.
NipponCapture began as a way to save useful Japanese from Safari without breaking the act of reading. It is now developing into a broader personal reading, typing, and study environment built from real language encounters.
The central idea remains simple: the Japanese a learner chooses to read is often more memorable and more personally relevant than material selected by a generic flashcard deck. NipponCapture preserves that material at the moment it appears, then gives the learner a place to review, organize, type, and revisit it later.
A capture can retain the selected text, its sentence and paragraph, the page title, source URL, and capture time. Once imported into the application, that material becomes part of a local library rather than disappearing into a notes file or an unstructured export folder.
Reading and Language Tools
The local application now includes a dedicated reading workspace, dictionary lookup, typing review, learner-progress views, and furigana support. Literary import and Aozora catalog tools extend the original Safari capture workflow.
Furigana preserves readings explicitly supplied by the source. Inferred readings use a separate language-analysis path, so an author's special or artistic reading is not silently treated as a general dictionary pronunciation.
Current development includes corpus preparation, source fidelity, and conservative reading evaluation. These are development capabilities; the application is not yet publicly released.
The Current System
Build A Personal Library
Captures are imported into a local library where duplicate records can be consolidated, deleted material can remain excluded from later cumulative imports, and passages can be reviewed as durable study records.
Practice By Typing
Library passages can become typing exercises that require the learner to reproduce Japanese text actively rather than only recognize it. Sessions record duration, correct and incorrect characters, skipped characters, accuracy, and typing speed.
Track Difficult Words
Words or text spans skipped during practice are stored with their passage and session, creating a record of where recall or reading fluency broke down and what may deserve later review.
Preserve Study History
Typing sessions and skipped words are stored in a persistent SQLite database so that practice can eventually be analyzed across passages, vocabulary, accuracy, and time.
Keep Data Portable
The application remains local-first and exportable. CSV workflows are being retained alongside the richer library and study database rather than locking the learner's reading history inside a closed service.
The Design Philosophy
“The learner's real reading life should be allowed to shape the curriculum.”
Many language-learning systems begin with a fixed syllabus and ask the learner to work through isolated words, example sentences, and graded exercises. NipponCapture begins from the opposite direction. The learner has already chosen something to read, has already encountered a gap in understanding, and has already shown that the subject matters enough to hold attention.
The goal is not to capture every unfamiliar word or turn browsing into constant homework. The goal is to make preservation cheap, defer judgment until later, and allow useful material to accumulate without interrupting the reading session that made it meaningful.
Typing is being used as an active reading exercise rather than as a speed game alone. Reproducing a passage forces attention onto kana, kanji, punctuation, word boundaries, and sentence structure. Skipping remains available because a learner should be able to move through difficulty without abandoning the session, while the skipped material is still recorded for later review.
The longer-term aim is to combine personal reading with carefully researched Japanese-language education materials. Future passage curation will examine what students in Japan actually encounter across elementary, middle, and high school: public curricula, textbook selections, test excerpts, library recommendations, dictionaries, and age-appropriate readers. The purpose is not to imitate an American assigned-reading list, but to understand the progression of Japanese literacy and use it to support learners at appropriate levels.
Why It Exists
Authentic reading creates a constant tension between immersion and study. Looking up every unknown word destroys momentum. Ignoring every difficulty preserves momentum but allows useful discoveries to disappear. Copying material across notes, dictionaries, spreadsheets, and flashcard tools creates friction and separates the language from its original context.
NipponCapture is intended to reduce that fragmentation. Capture should take only a moment. Organization, typing, explanation, tagging, export, and deeper review should happen later, outside the original reading session.
Over time, the library can become a record of what the learner actually encounters: recurring vocabulary, difficult kanji, skipped words, preferred subjects, frequently read sources, typing accuracy, and passages that repeatedly demand attention. That history can support a study system that is personal without becoming arbitrary.
Current Progress
| Area | Current State |
|---|---|
| Safari capture | Working context-menu capture of selected Japanese into shared local storage. |
| Captured context | Stores selected text, sentence, paragraph, page title, URL, capture date, and archive information. |
| Library import | Working import into the macOS library with capture-ID deduplication and consolidation of matching source URLs and sentences. |
| Deletion protection | Deleted capture IDs are recorded so later cumulative CSV imports do not restore material the user intentionally removed. |
| Typing practice | Working passage-based typing sessions with cursor progression, correctness tracking, skip behavior, and synchronized passage rendering under active refinement. |
| Session persistence | Working SQLite storage for duration, correct characters, incorrect characters, skipped characters, accuracy, characters per minute, and completion state. |
| Skipped-word history | Working persistence of skipped text spans with the original surface form, passage, character range, session, and creation time. |
| Reading workspace | Reader, context and dictionary inspectors, typing review, and learner-progress views are present in the local application. |
| Furigana and reading analysis | Explicit source readings are preserved separately from inferred lexical readings. Work continues on source fidelity and conservative reading evaluation. |
| Literary sources | Aozora import and catalog tools support corpus preparation, provenance, and human review. Catalog flags are treated as review evidence, not blanket reuse permission. |
| Public availability | Not currently released. The project remains a local experimental application under active development. |
Development Path
- FOUNDATIONSafari capture, local library import, deduplication, deletion protection, persistent typing sessions, and skipped-word history.
- CURRENTReading and dictionary workspaces, source-aware furigana, learner-progress tools, and literary corpus preparation and evaluation.
- NEXTContinue validating reading quality and source fidelity, refine the connected reading and practice experience, and develop curated study material before a public release.