#3535th of 39 in RAG, knowledge and searchFor science, health and research, research and papers
local-deep-researcher
Iterative web research agent that runs on Ollama or LM Studio models
- Stars
- 9.4k
- License
- MIT
- Last commit
- Oct 2026
- Language
- Python
8,889 on 1 Apr 20269,386 on 21 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
LangChain is hiring: 104 open roles
Overview
Takes a topic, has a local LLM write a search query, summarizes the results, reflects on knowledge gaps and searches again for a configurable number of loops (default 3). The output is a markdown summary with cited sources, stored in LangGraph state. It runs as a LangGraph graph that you inspect through LangGraph Studio.
Who it is for: Self-hosters wanting local-LLM web research reports via LangGraph
Strengths
- DuckDuckGo search by default, so no API key is needed to start
- Supports Ollama and LM Studio, with SearXNG, Tavily and Perplexity as search options
- Final markdown report cites all sources gathered during the loops
- Fallback handling for models that fail structured JSON output
Weaknesses
- No standalone UI; interaction goes through LangGraph Studio, opened via smith.langchain.com
- Small models such as DeepSeek R1 1.5B and 7B struggle with required JSON output
- Docker image does not include Ollama; it must be run and configured separately
- Safari shows mixed-content warnings with Studio; Firefox is recommended
What it needs
- GPU optional
- Docker
- Needs Ollama or LM Studio, LangGraph CLI, uv, DuckDuckGo, SearXNG, Tavily or Perplexity for search
- Models: Ollama models (default llama3.2), LM Studio OpenAI-compatible server, DeepSeek R1 distills, qwen_qwq-32b
- port 2024
Also in RAG, knowledge and search
See all 39| Rank | Project | Score |
|---|---|---|
| 1 | deeptutorSelf-hosted AI tutor with knowledge bases, guided learning, books and quizzes | 82 out of 100 |
| 2 | LightRAGGraph-based RAG framework with a REST API server and web UI | 78 out of 100 |
| 3 | MorphicAI answer engine with cited search results and generative UI components | 77 out of 100 |
| #4 | MaxKBEnterprise knowledge-base agent platform with RAG, workflows and MCP tools | 75 out of 100 |
| #5 | RAGFlowRAG engine with deep document parsing, template chunking and agentic retrieval | 74 out of 100 |