Our Roadmap

Access to the calculator and calendar as a tool for Apple Intelligence to enable accurate calculation of numerical data and extraction of dates from documents. The tool's settings have also been improved.

At the tap of a button, the app generates an AI-powered summary of all documents in a project. Users choose the level of detail upfront: Brief (executive summary, ~3 sentences), Balanced (key points + sources, ~1 page), or Detailed (full analysis with outline, multiple pages). The result can be exported as Markdown or PDF.

Once a file has been indexed, the on-device AI automatically generates three basic tags and a short summary. This process is implemented as a separate queue, allowing search queries to be executed immediately.

Spreadsheet data can be imported as a RAG source. Row-based chunking prefixes each chunk with the column header. Questions like "Which quarter had the highest revenue?" become possible.

Presentations can be imported as a RAG source. Parsing is handled via ZIPFoundation (already a project dependency), mirroring the existing DOCX parser.

A small or medium widget displays the most recently used project and opens the chat directly via deep link. It increases daily active usage without requiring a full app launch.

Core actions such as querying a project, listing projects, and opening a project are registered as App Intents. They are discoverable in Spotlight and callable via Siri — a powerful differentiator on iOS 26.

Project names and document titles become searchable via iOS Spotlight. A tap opens the correct view in the app directly.

The AI generates a structured script from project content; the on-device text-to-speech engine reads it aloud. Fully offline and private — a direct differentiator over NotebookLM, which relies on cloud processing.

Highlights and sticky notes directly inside the app's document viewer, without switching to an external app. Annotations are stored in the database as JSON alongside the document record — not a full PDF editor.

Two versions of a document (e.g. contract v1 vs. v2) are compared side by side. Changes are visually highlighted and summarised by AI. This premium feature is planned for the Pro+ tier.

Implementation of the Personal Voice text-to-speech feature for devices with English as the language setting. The modern Personal Voice will be implemented as an optional alternative to the standard text-to-speech feature.

Video import via on-device transcription and indexing of the transcription.

Automatic on-device transcription. The transcript is then indexed for search.