PrivyPDF

Overview

PrivyPDF is a local-first macOS app that turns your PDFs into a searchable, question-answerable knowledge base. All processing runs on your Mac. Your documents and embeddings stay in your App Sandbox; no data is sent to the cloud.

How it works

• Import — Add PDFs via Import or drag-and-drop
• Parse & Index — Extract text and build a local vector index using your own Ollama server
• Ask — Type questions in natural language and get answers with source citations
• Jump — Tap a citation to jump to the exact page (and region) in the PDF preview
• Export — Export document text as Markdown or JSON for use elsewhere

Privacy by design

PDFs and indexes are stored only in the app's sandbox. AI runs via Ollama on http://localhost:11434. The app does not ship or download model weights. No account, no telemetry, no external servers for your documents.

Requirements

• macOS 14.6 or later
• Ollama — Installed and running locally, with at least one embedding model (e.g. nomic-embed-text) and one chat model (e.g. qwen2.5:7b-instruct). Configure the base URL and models in PrivyPDF → Settings. Download at ollama.ai

Who it's for

• Developers and engineers who want to query technical docs offline
• Researchers and knowledge workers who need citations and local control
• Anyone who values privacy and wants a PDF copilot that never uploads their data