Player not loading? Watch on YouTube
This tutorial sets up Local Deep Research with Ollama and SearXNG through Docker Compose on Ubuntu. The presenter uses an NVIDIA RTX A6000 with 48 GB of VRAM, but says smaller models can use less memory.
The walkthrough covers cloning the repository, starting the containers and pulling Gemma 7B into Ollama. A model download initially fails because the command uses the wrong container name. The presenter checks it with Docker PS and retries. He also encounters a SearXNG restart loop, attributes it to permissions, adjusts them and restarts the service. The browser interface runs at localhost port 5000 and requires account registration.
The presenter describes a self-hosted research workflow that breaks questions into subquestions, searches web and academic sources, and indexes collected material in an encrypted local library.
The example asks about premature aging using quick-summary mode. The interface also offers detailed reports and local PDF uploads. An apparent completion state changes back to active research; after waiting, the presenter shows a report with citations and PDF export.
