Open-source Deep Research

Deep Research for LlamaIndex

Add broad cited discovery around LlamaIndex apps without tying research to one model.

Works beside indexed knowledge

LlamaIndex is excellent for structured retrieval over known corpora. AutoSearch complements that by finding external evidence from public channels, developer communities, academic sources, and Chinese platforms when the answer is not already in your indexed data.

MCP keeps integration explicit

AutoSearch exposes research actions through MCP, so teams can connect it as a tool rather than merge source discovery into every index pipeline. That boundary makes experiments easier and keeps private data retrieval separate from public research.

Citations for research outputs

When LlamaIndex apps produce reports or analyst notes, source traceability matters. AutoSearch returns cited findings from multiple channels, giving downstream synthesis steps clearer provenance for claims, comparisons, and recommendations.

How it fits

AutoSearch sits beside LlamaIndex as a public research companion. LlamaIndex can continue serving private documents, knowledge graphs, and domain indexes. AutoSearch is called when the workflow needs current external context from 40 channels, including 10+ Chinese sources. The two systems stay distinct: LlamaIndex organizes known data, while AutoSearch gathers cited outside evidence for the agent or application to synthesize.

Try this prompt

Research current examples of LlamaIndex agents using external tools.
Compare docs, repositories, and recent discussions with citations.