BigHugger
GH Repository · zilliztech

deep-searcher

Open Source Deep Research Alternative to Reason and Search on Private Data. Written in Python.

stars
8,268
30-day movement
+10/day
Related entries
60
Connections
1
agent-appPythonqwen3claudereasoning-modelsagentpythonllmgrokragmakedockerllama4vector-databaseagentic-ragdeepseekzillizmilvusgrok3openaideep-researchpython/uvdeepseek-r1

deep-searcher is a Python open-source alternative to deep research tools that reasons over and searches private data, described as an agentic RAG application. The repository ships a deep-searcher package with Docker, Make, and Python uv toolchain support, and its topics indicate integration with vector databases (Milvus/Zilliz) and multiple LLM providers (DeepSeek, OpenAI, Claude, Qwen, Grok, Llama).

Use it when you want deep-research-style reasoning and search over your own private data rather than the public web.

Use it to

  • Run deep research queries against private documents
  • Build agentic RAG pipelines with Milvus or Zilliz
  • Swap between reasoning LLM providers like DeepSeek-R1, Claude, or Qwen
  • Deploy locally via the provided Docker setup

For Developers building search and research agents over private data

Role
agent-app
Language
Python
Licence
Apache-2.0
Forks
803
Open issues
42
Last push
2025-11-19
Latest release
embedding_model · 2025-02-20
topicsagentic-ragdeep-researchvector-databasellmmilvuspython