BookJumperFrom your first book to a shared library
Reading guide
Open a document, organize it, annotate it, and ask questions. Learn what each step does and how to recover when something needs attention.
01 · Open your first document
- Sign in, open the library, and create a subject folder such as Economics or Dissertation.
- Choose Open local file, select a document and its subject, and label it as a book, handout, or paper.
- Local reading is the default, including study rooms. Uploading and indexing start only when you explicitly select synchronization. OCR still needs a separate page selection and confirmation. A failed upload does not interrupt the original document.
- Open the document from its subject folder. PDF and images display the original; Office documents, ebooks, and text use a readable text view.
02 · Read, annotate, and compare
- Use the contents, page controls, and search to navigate. PDF page numbers match physical pages; text and ebooks are paginated for reading and may differ from a printed edition.
- Select text to highlight or annotate it with its source position. Split reading helps compare books, handouts, and notes.
- The original preserves the source layout; the text index enables search and questions. They can finish at different times. Wait for the index to update after replacing a version.
03 · Choose pages to recognize scan text
- OCR turns image text into searchable content. Opening a local PDF does not extract all its text. Explicit synchronization enables native text processing; importing, opening, synchronizing, searching, and asking AI never starts OCR automatically.
- Only PDFs, regular images, and multipage TIFFs support page OCR. Open Page OCR in the document, select or enter exact pages such as 3-6, 12, review the count, and confirm. Only confirmed pages enter the queue.
- Study rooms reuse results by shared material, file version, and page: if one member has completed or is processing a page, another member selecting it does not create another OCR run. Select pages again for a replacement file version. Members with reading access and authorized share-link visitors can view completed results; only signed-in active members can start OCR or AI.
- DOCX, PPTX, and EPUB extract native text only. Their embedded scans are never sent to OCR automatically. Export the relevant pages to PDF, then select the PDF pages; retrying also processes only failed pages you explicitly select again.
- Each file is limited to 25 MB, 1,500 pages, and 2.4 million characters. Spreadsheets have additional limits. Split oversized files by chapter or worksheet; renaming the extension does not convert them.
04 · Share a study room
- Choose Join / Create study room in the library. Its creator becomes an administrator and receives a 12-character alphanumeric invitation code.
- Join by code or share a link. Members can invite others, create subject folders, and add shared material.
- Link visitors can read shared documents without signing in. AI, contributions, and saving the room require sign-in; you return to the original room afterward.
- Administrators can remove members and members can leave. Follow the administrator handover prompt when the last administrator leaves. Shared links allow guest reading, so removing a member does not revoke circulated links or downloaded files.
- Personal library and room materials have different visibility. Check the destination and sharing description before adding a document.
05 · Maintain a room timetable
- Open Study room timetable from a room or the reader calendar action. Day, week, month, Gantt, agenda, and analytics show the same shared schedule.
- Import ICS, Excel (XLSX / XLS), ODS, CSV / TSV, JSON, MBA HTML, or text, or paste a dated course notice. Parsing and preview run locally without a paid model call. Review warnings and confirm before saving to the room.
- Maintain courses and individual class times; use course details for exams and contacts. Settings include title, time zone, teaching week, and breaks.
- Active members can edit by default. The creator can revoke individual editing permission and transfer ownership. History records the actor, timestamp, and actual before-and-after changes.
- Tasks, notes, attendance, ratings, and enrollment preferences stay on this device. If edits conflict, use Compare and merge. Independent field edits are retained; overlapping changes require an explicit choice and confirmation. A further conflict requires a fresh review. You can also export the unsaved draft. Shared JSON backups and ICS exports are available.
06 · Ask and organize with sources
- Once text is indexed, ask about the current book, chapter, or subject. Be specific: for example, compare the definitions of opportunity cost on pages 3 and 8.
- Follow citations back to the original. Recognition quality, source coverage, and model output affect answers; scan pages you have not explicitly selected for OCR, or that are still processing, are not OCRed by a question and may not yet be included.
- Use mind maps, knowledge relationships, content maps, and quizzes to organize reading. Author and character conversations are AI interpretations grounded in the text.
- Read-aloud, transcription, and live voice require their configured services and permissions. Grant microphone access before recording. Provider errors can be retried later.
Supported documents
| Format | Reading behavior and limitations |
|---|---|
| Original pages and page numbers; select exact scanned pages and confirm before OCR starts. | |
| DOCX / DOC | DOCX text opens locally. Legacy binary DOC keeps its original and requires conversion to DOCX/PDF or explicit cloud sync. Embedded scans are not OCRed automatically. |
| PPTX / ODP | Native slide text in presentation order, including blank slides. Export scanned slides to PDF before page-by-page OCR. |
| XLSX / XLS / ODS | Read displayed values by worksheet; formulas are not recalculated and macros are not executed. |
| CSV / TSV | Read delimited tables, quoted line breaks, and identifiers with leading zeros. |
| EPUB | Read native chapter text in publication order; export scanned chapters to PDF for OCR. |
| TXT / MD / HTML / RTF / ODT | Readable text with common Chinese encodings; HTML scripts and style content are removed. |
| PNG / JPG / WebP / GIF / BMP / AVIF / TIFF | View original images, including multipage TIFF. OCR starts only after you select an image or page and confirm; animated images use the first frame. |
Text reading does not reproduce every feature of the original application. Export to PDF when layout, equations, or charts must match page by page. Keep the original for comparison. Convert legacy PPT and HEIC / HEIF first; unlock password-protected or DRM-protected files legally in their original application.
Troubleshooting
The file opened, but I cannot see text
Local reading does not depend on uploading or AI processing. PDFs and scans open as originals; restricted storage is labeled as available for this session only. Opening a file never starts OCR automatically. To make scan text searchable, choose the exact pages in the document and confirm OCR. For damaged, encrypted, or invalid files, reopen and export from the original application; changing the extension does not convert a file.
Why can AI not find content in a scanned PDF?
Seeing the original page confirms the file is open, not that its scan text is indexed or that OCR will start automatically. Open Page OCR in the document, select pages such as 3–6 and 12, then confirm; only those pages are processed. In a study room, completed or in-progress pages from the same file version are reused. Small print, handwriting, blur, rotation, and complex formulas can reduce accuracy; check the original page.
Why is my file missing on another device?
Local reading and cloud synchronization have separate states. Confirm synchronization completed and use the same account in the same region. Keep the window open while uploading. If synchronization fails, keep the local original and retry.
The WeChat code is slow, expired, or did not finish
Refresh the QR code, then confirm in WeChat while keeping the original window open. Refreshing invalidates the earlier attempt. Email, username, and phone sign-in remain available. A first-time WeChat identity needs an explicit account creation or binding step.
Can I keep reading when AI or voice is unavailable?
Documents that are already open do not depend on every AI request succeeding. Check connectivity, sign-in, and indexing status. Recording and live voice also need microphone permission. Retry later when a provider times out; avoid repeated submissions.
How do I keep the documents in my existing account?
Sign in with your existing email, username, or phone number and bind WeChat in account settings. Creating another account does not move your documents. Global and China regions store data separately; a WeChat identity does not automatically transfer books or notes between them.