https://yanghu.github.io/tags/ollama/Recent content in Ollama on Yang's NotesHugo -- gohugo.ioen-usyang@yhu.me (Yang Hu)yang@yhu.me (Yang Hu)© 2026 Yang HuThu, 12 Mar 2026 00:00:00 -0800https://yanghu.github.io/posts/paperless-ai-setup/Thu, 12 Mar 2026 00:00:00 -0800yang@yhu.me (Yang Hu)https://yanghu.github.io/posts/paperless-ai-setup/<p>This post is a complete runbook for integrating AI-powered auto-tagging and classification into <a href="https://docs.paperless-ngx.com/" target="_blank" rel="noreferrer">paperless-ngx</a> using <a href="https://github.com/clusterzx/paperless-ai" target="_blank" rel="noreferrer">paperless-ai</a> and a locally-running <a href="https://ollama.com/" target="_blank" rel="noreferrer">Ollama</a> instance. The setup uses a local LLM to read document text and automatically populate metadata fields — title, document type, tags, correspondent, date, and custom fields.</p> <h2 class="relative group">Hardware and Architecture <div id="hardware-and-architecture" class="anchor"></div> <span class="absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100 select-none"> <a class="text-primary-300 dark:text-neutral-700 !no-underline" href="#hardware-and-architecture" aria-label="Anchor">#</a> </span> </h2> <ul> <li><strong>NAS (Synology DS1621+, <code>10.0.10.10</code>)</strong>: runs paperless-ngx on port 5656</li> <li><strong>Desktop PC</strong>: Windows with WSL2, Docker Desktop, RTX 4090</li> <li><strong>Goal</strong>: AI auto-tagging/classification using a local LLM, zero cloud dependency</li> </ul> <p>The key architecture decision is a <strong>pull model</strong>: paperless-ai runs in WSL2 Docker, polls the paperless-ngx API for documents tagged <code>ai-pending</code>, processes them with Ollama, and writes metadata back. This is the correct approach for a desktop that is not on 24/7 — the NAS holds the queue and the desktop drains it when available.</p>