AI PDF Reader v2 is here: ask any PDF, answered privately on your device

Today we're releasing AI PDF Reader v2 — a ground-up rebuild of our PDF question-answering app on the cross-platform AI Suite v2 stack. The idea is the same one we shipped in v1: open a PDF and ask questions about it, answered locally. Everything around that idea is new.

What it does

Open any PDF and AI PDF Reader indexes it on your own device. Ask a question and the answer streams back with inline [1][2] citations and a sources row — click either to jump to the exact page with the passage highlighted, so you can verify every answer against the document itself. With a local model, extraction, indexing, retrieval and answering all run on your machine; turn the network off and it keeps working.

It's also a real PDF reader: zoom and fit, rotate, continuous or single-page layouts, a thumbnail sidebar, find-in-document, text selection and copy, and hands-free auto-scroll with a reading-speed tracker. You can use it purely as a reader even if you never ask a question.

New in v2

• Page-exact citations. Every answer points to the page it came from — auditable, not a black box.
• Annotations and bookmarks that travel. Highlight, underline, strike through, add notes and shapes, and bookmark pages with labels, colours or a custom image. Annotations and bookmarks are written into the PDF itself, so they come along when you share the file — and a private encrypted copy stays on your device.
• Your choice of models. Start free with a capable on-device model (selected and downloaded automatically on your first question). Pro unlocks premium and multilingual models, and optional cloud providers for the hardest questions.
• Pro power tools. OCR makes scanned, image-only PDFs answerable; whole-document summaries; and "ask across every open PDF at once," with citations that name the file and page.
• In 16 languages. The whole interface — viewer, panels and menus — follows your operating system's language.
• Cross-platform and scriptable. Windows, macOS and Linux, plus a headless command-line tool that doubles as an MCP server so AI agents can query your PDFs on your own infrastructure.

Why on-device matters

Contracts, case files, patient records, financial reports — the documents people most want to ask questions about are often the ones that can't go to the cloud. AI PDF Reader brings the AI to the documents instead: nothing leaves the device with a local model, all of your data is encrypted at rest, and there are no per-page or per-token usage fees. It's the same on-premises, privacy-first principle behind everything Software Tailor builds ^[1].

Get it

AI PDF Reader v2 is available now on the Microsoft Store for Windows as a free download, with Pro available for power features ^[2]. macOS and Linux desktop builds are available too, and browser and mobile editions are on the roadmap. Read the full feature tour on the product page.

References

1. Software Tailor. "AI PDF Reader — product page." softwaretailor.com/ai-pdf-reader.htm. Accessed 2026-06-16.
2. Microsoft. "AI PDF Reader on the Microsoft Store." apps.microsoft.com/detail/9NP5SNCVS46V. Accessed 2026-06-16.

Related articles

• AI Admin Console: what each of the six capabilities actually does
• Content-free audit logs: how we prove what the model did without keeping prompts
• Why on-prem AI deployments stall in procurement (and the framing that fixes it)
• All articles →