Originally published on andrew.ooo — visit the original for any updates, code snippets that aged out, or follow-up posts.

TL;DR

Local Deep Research (LDR) is an open-source AI research assistant from LearningCircuit that does what ChatGPT's "Deep Research" and Perplexity Pro do — but on your own hardware, against your own LLM, with your own search backend, and with everything stored in an AES-256 encrypted SQLite database that even the server admin can't read.

Key facts as of early May 2026:

~5,950 GitHub stars , ~1,100 added this week — currently on GitHub's weekly Python trending list

~95% accuracy on SimpleQA (preliminary, GPT-4.1-mini + SearXNG + focused-iteration strategy) — broadly comparable to closed-source deep research products

Apache-2.0 licensed, packaged on PyPI as local-deep-research , with signed Docker images on Docker Hub

20+ research strategies including a new langgraph-agent mode where the LLM decides which engines to use and when to synthesize

10+ search engines out of the box:…