PDF-Pals

PDF-Pals is a PDF solver tool. A native Mac app for chatting with PDFs, including scanned files via OCR, with answers tied to cited source pages. Documents stay on the local disk, and users can bring their own OpenAI, Anthropic, Google or OpenRouter key, or run local models through Ollama and LM Studio.

PDF Pals is a native macOS application for reading and interrogating PDF documents with an AI assistant. You open one or several PDFs, ask questions in ordinary language, and receive answers that point back to the specific source pages supporting them. Because it is built as a native Mac app rather than a wrapped web interface, it handles large files without the file size ceiling common to browser based document chat tools. The privacy model is the defining design choice. Documents are never uploaded, the library lives on the local disk, and no cloud vector database is involved. Indexing happens on the machine, and the data sits in a local SQLite database that can be backed up like any other file. API keys are kept in the Apple Keychain, and crash reports plus analytics are optional. Behind those settings sit the model providers: PDF Pals can use OpenAI, Anthropic, Google, OpenRouter, Ollama or LM Studio, so teams that already hold an API key can bring it, and users who prefer offline processing can point the app at a local model instead. The homepage is explicit that no account is required to start. In practice the app is aimed at people who read dense material for a living. Researchers use it to summarize papers and compare findings across several documents. Legal professionals ask questions of contracts and case files, then follow each answer to the page it came from. Developers work through specifications and technical books by jumping straight to the relevant passage. A built in OCR engine extends that to scanned PDFs and complex forms that would otherwise be unsearchable, and the system prompt and model settings can be adjusted for repeatable workflows. The tradeoff is scope. PDF Pals is a Mac only desktop product, with system requirements of macOS 13 or later for the current release and macOS 14 or later for version 2, and there is no Windows or Linux build. Licensing is a one time purchase with a defined period of updates rather than a subscription, which suits buyers who dislike recurring fees but means ongoing updates are a separate consideration. Against browser based PDF chat services, the pitch is speed, no upload step and verifiable citations.