Skip to content

Local Deep Research

by LearningCircuitAgentResearch something

Researches a question across the web, papers and your own files and writes a cited report

View on GitHub

About Local Deep Research

Local Deep Research is an open source research assistant you run on your own computer and use in the browser. You ask a question; it searches the web, papers and your files, then writes a cited summary or report.

For Students and researchersPeople who prefer local AIAnyone writing cited reports

Made by LearningCircuit

An open-source project published on GitHub under the account LearningCircuit, which names no person or company. It is released under the MIT licence.

Stars
★ 9.5k
License
MIT
Cost
Free
Platforms
Mac · Win · Linux
Level
Advanced

Overview

Local Deep Research is a research assistant you run on your own computer and use in the browser. You ask a question; it decides what to search across web, academic and local sources, then writes a quick summary, a detailed analysis or a full report with citations. It can save the sources it finds into an encrypted personal library and search your own documents next time. It suits students, researchers and anyone who wants deep research with local models.

A local model can do the thinking on your own hardware, and each user's data stays in an encrypted database with no telemetry.

Try these first

Start with the first one: it is a small task you can check easily.

  1. Quick summary: [a question you already know the answer to]. List every source you used.First try · searches the web

  2. Write a detailed report on [topic] using arXiv and PubMed, with a table of contents.searches the web

  3. With the scope set to Private only, ask: what do my saved papers say about [topic]?

  4. Subscribe to [topic] with a weekly schedule and summarise only new developments.runs on a schedule

Example prompts written by Agent Flows Hub.

Features

From the creator's documentation.

FeatureWhat it doesRuns
Research modesQuick summary, detailed research and full reports with sections and citationsYour chosen model
Search sourcesWeb, arXiv, PubMed, Semantic Scholar, Wikipedia and more; paid engines are optionalInternet
Personal librarySave found sources to an encrypted library and search your own documentsYour computer
Privacy scopesChoose per run whether public engines and cloud models may be usedYour computer
Export and historyDownload results as PDF or Markdown and revisit past researchYour computer
News digestsScheduled research on topics you subscribe toInternet

Good to know

Ask a research question in the browser and choose a quick summary or a full report. It searches the sources you allow, then writes an answer with citations you can open and check.

Limits

  • You must pick a model in Settings before the first run. A local model needs enough hardware; a cloud model sends your prompts to that provider.
  • Web searches still go to the internet, even with a local model. Only the Private only scope limits a run to your own collections, and the README says it is not a hard boundary.

Notes by (original project name: Local Deep Research (LDR)), written from the creator's documentation, not from running it ourselves. Local Deep Research README · Local Deep Research installation guide

Stuck setting it up?

Tell us what happened. We take a look first; if it needs paid work, you get an estimate before anything starts.

Get help

Other agents

  • Lovable Builds and publishes web apps and websites from a description in plain words
  • Manus Takes a one-line request and delivers finished slides, websites, research and reports
  • Mercury Agent Keeps a personal AI agent running on your computer that you can message from Telegram
  • MiMo Code Lets you hand coding tasks to Xiaomi's terminal agent, which remembers your project
  • nanobot Runs a personal AI agent in your browser or chat apps that works with files and schedules
  • Nanobrowser Works through web pages for you in Chrome, clicking, typing and moving between tabs

See all 61 agents · Install your first AI agent

GitHub