Readwise Reader

Aggregates articles, newsletters, PDFs, tweets, YouTube videos, RSS feeds, and EPUBs into a single web-based reading application accessible at readwise.io/read. Uses a centralized document store with metadata tagging and source attribution, eliminating the need to switch between Pocket, Instapaper, email clients, and social media platforms. Content is indexed for full-text search and organized via user-defined tags and collections.

Enables users to ask natural language questions against the full text of saved documents (articles, PDFs, newsletters, transcripts) using GPT-4 as the underlying LLM. The system passes document content as context to GPT-4 and returns answers grounded in that specific document. Implementation details (context window size, token limits, error handling) are undocumented, but the feature operates on a per-document basis rather than cross-document search.

Automatically extracts transcripts from YouTube videos when a video URL is saved to Reader. Transcripts are indexed for full-text search and support the same highlighting and annotation features as articles and PDFs. Feature enables searching within video content and creating highlights from transcript text. Transcript availability depends on YouTube's caption availability; auto-generated captions may be used if manual transcripts are unavailable.

Enables users to subscribe to RSS feeds and automatically aggregate new articles into Reader. Subscribed feeds are polled on a regular schedule (frequency not documented) and new articles are added to the reading queue. Feed management (add, remove, organize by category) is provided through the Reader interface. Articles from RSS feeds are treated identically to manually saved articles, supporting the same highlighting, tagging, and export features.

Generates summaries of saved content (articles, PDFs, newsletters) using an unspecified AI model (claimed as 'AI-powered' but model identity not documented). Summarization trigger (automatic vs. on-demand), length parameters, and caching behavior are undocumented. Feature appears to operate on individual documents and is presented as part of the Reader feature set, but technical implementation details are absent from public documentation.

Browser extension enables one-click saving of web articles directly to Readwise Reader from any webpage. Provides in-page highlighting and annotation overlay that persists with saved content. Extension integrates with the browser's native UI (likely via sidebar or context menu) and syncs highlights back to the centralized Reader application. Specific browser support (Chrome, Firefox, Safari, Edge) and keyboard shortcuts are undocumented.

Indexes all saved content (articles, PDFs, newsletters, transcripts) and provides full-text search capability accessible from the Reader interface. Search operates across document bodies, titles, and user-created tags. Implementation approach (inverted index, vector embeddings, or keyword matching) is undocumented. No indication of AI-augmented semantic search or relevance ranking beyond basic keyword matching.

Integrates with spaced repetition systems (implied to include Anki, SuperMemory, or similar) to resurface saved highlights and notes on a configurable schedule. Daily review can be delivered via email or accessed through the Reader app interface. Integration mechanism (API, export format, or direct sync) is undocumented. Feature appears to operate on user-created highlights rather than auto-generated summaries.

Enables export of saved articles, highlights, and annotations to external note-taking platforms (Notion, Obsidian, Evernote). Export mechanism (API integration, webhook, or scheduled sync) is undocumented. Feature appears to support one-way export rather than bidirectional sync. Export format and customization options (which fields to include, formatting) are not documented.

Provides a tagging system allowing users to organize saved content with custom tags and create collections. Tags are searchable and filterable within the Reader interface. Implementation uses a flat tag structure (no hierarchical categories documented). Tags persist with documents and are included in exports to external tools. No automated tagging or tag suggestions documented.

Synchronizes saved articles, highlights, annotations, and reading progress across devices (web, mobile, desktop) through a centralized cloud backend. Sync mechanism (real-time, eventual consistency, or scheduled) is undocumented. Feature enables starting reading on one device and resuming on another without manual state transfer. Offline support and conflict resolution strategies are not documented.

Enables users to upload PDF and EPUB files directly to Reader for annotation, highlighting, and searching. Files are stored in the centralized Reader backend and treated as first-class content alongside web articles and newsletters. Uploaded files are indexed for full-text search and support the same highlighting and annotation features as web content. File size limits and storage quotas are not documented.