Repolode
投稿する
LearningCircuit

LearningCircuit/local-deep-research

Self-hosted AI research assistant that runs deep, multi-step searches across local and cloud LLMs and many search engines, with citations and encrypted storage.

代替できる製品:PerplexityChatGPT Deep Research

GitHub で見る
スター
9,331
フォーク
869
未解決の Issue
1,253
最終コミット
5 時間前

これは何？

Local Deep Research is a self-hosted research assistant. Given a question, it searches the web, academic databases such as arXiv and PubMed, and your own documents, then synthesizes the findings into a report with citations. It works with local models served by Ollama, LM Studio or llama.cpp, as well as cloud providers such as OpenAI, Anthropic, Google Gemini, OpenRouter and Requesty.

The project offers quick summaries, detailed research and full report generation, plus a LangGraph-based agent strategy in which the LLM decides which engines to query and when to stop. Downloaded sources can be indexed into a personal encrypted library and searched alongside live web results. Each user's data is kept in an SQLCipher database encrypted with a key derived from their password.

LDR also exposes a Python API, an HTTP API, an MCP server for Claude Desktop and Claude Code, and LangChain retriever integration. An egress scope setting controls which search engines and model providers a run may use, including a private-only mode that forces local inference.

主な機能

  • 01Multi-step agentic research that picks search engines and synthesizes cited reports
  • 02Supports Ollama, LM Studio, llama.cpp, OpenAI, Anthropic, Gemini and OpenRouter models
  • 03Searches arXiv, PubMed, Semantic Scholar, Wikipedia, SearXNG and other engines
  • 04Indexes downloaded sources and your own documents into an encrypted local library
  • 05Per-user SQLCipher databases with encryption keyed from the user's password
  • 06Python API, HTTP API and MCP server for Claude integration

ユースケース

  • Literature reviews that need searches across arXiv, PubMed and the open web with citations
  • Querying confidential documents while keeping inference and data on your own hardware
  • Building a searchable personal research library that grows with each research session
  • Automating recurring topic digests from academic and news sources

トピック

#ai#deep-research#rag#local-llm#ollama#self-hosted#search#academic-research#academia#anthropic

よくある質問

local-deep-research とは？
Self-hosted AI research assistant that runs deep, multi-step searches across local and cloud LLMs and many search engines, with citations and encrypted storage. Local Deep Research is a self-hosted research assistant. Given a question, it searches the web, academic databases such as arXiv and PubMed, and your own documents, then synthesizes the findings into a report with citations. It works with local models served by Ollama, LM Studio or llama.cpp, as well as cloud providers such as OpenAI, Anthropic, Google Gemini, OpenRouter and Requesty.
local-deep-research は無料で使えますか？ライセンスは？
local-deep-research は MIT ライセンスのオープンソースです。商用製品で使う前にライセンス条件を確認してください。
local-deep-research は今もメンテナンスされていますか？
活発にメンテナンス中。local-deep-research の最終コミットは 2026年10月19日 です。
local-deep-research は何に使えますか？
Literature reviews that need searches across arXiv, PubMed and the open web with citations Querying confidential documents while keeping inference and data on your own hardware Building a searchable personal research library that grows with each research session Automating recurring topic digests from academic and news sources
local-deep-research はどこで入手できますか？
ソースコードとインストール手順は GitHub にあります：https://github.com/LearningCircuit/local-deep-research

概要はプロジェクトの README をもとに AI が生成しています。GitHub の数値は毎日更新されます。

関連プロジェクト