# ZeroRAG: Adaptive Semantic Peer Routing for Decentralized Retrieval-Augmented Generation

DevFeed: [ZeroRAG: Adaptive Semantic Peer Routing for Decentralized Retrieval-Augmented Generation](<https://devfeed.tech/articles/zerorag-adaptive-semantic-peer-routing-for-decentralized-retrieval-augmented-generation-120269.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.07928>)

Author: Petru Neague, Quinten Stokkink, Martijn de Vos, Rafael Pires, Mathis Randl, Anne-Marie Kermarrec, Johan Pouwelse

Published: 2026-11-06T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Distributed Computing](<https://devfeed.tech/sources/arxiv-distributed-computing.md>), [arXiv: Networking](<https://devfeed.tech/sources/arxiv-networking.md>)

Topics: [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [routing-system](<https://devfeed.tech/topics/routing-system.md>), [AI-driven information retrieval](<https://devfeed.tech/topics/ai-driven-information-retrieval.md>), [Vectors & LLMs](<https://devfeed.tech/topics/vectors-llms.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-dc](<https://devfeed.tech/tags/cs-dc.md>), [cs-ir](<https://devfeed.tech/tags/cs-ir.md>), [cs-ni](<https://devfeed.tech/tags/cs-ni.md>), [decentralized](<https://devfeed.tech/tags/decentralized.md>), [generation](<https://devfeed.tech/tags/generation.md>), [peer](<https://devfeed.tech/tags/peer.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [retrieval-augmented-generation](<https://devfeed.tech/tags/retrieval-augmented-generation.md>), [routing](<https://devfeed.tech/tags/routing.md>), [semantic](<https://devfeed.tech/tags/semantic.md>)

## AI overview

ZeroRAG routes RAG queries across peers while documents remain with their owners, avoiding centralized orchestration and globally coordinated data structures. Each peer keeps a bounded set of neighbor advertisements and a cache of recent queries; a best-first walk uses this local state to route queries and gradually rewire the overlay toward semantically relevant peers. The abstract reports tests on three real-world workloads against two decentralized baselines: ZeroRAG improved recall while requiring fewer network messages, approached a centralized store's recall and LLM answer accuracy, recovered after catastrophic peer failures, and scaled to 100,000 peers.

## Source excerpt

arXiv:2611.07928v1 Announce Type: cross Abstract: Retrieval-augmented generation (RAG) is a standard approach to enhance the reliability of large language model (LLM) outputs by retrieving documents relevant to a query and supplying them in the LLM prompt. Today's RAG systems collect, index, and serve these documents from centralized stores. In practice, however, documents are owned by organizations, devices, and individuals who cannot or will not surrender custody of them to a central party, e.g., because of privacy, copyright or regulatory compliance. Existing modern decentralized retrieval techniques either reintroduce a coordinating authority, or rely on a global data structure that requires continuous network-wide coordination (e.g., DHTs). We present ZeroRAG, a decentralized routing layer for document retrieval in RAG that performs semantic query routing across an overlay of peers with- out global data structures or centralized orchestration. Documents remain with the peers that own them, and each peer keeps only lightweight local state: a bounded set of neighbor advertisements and a cache of recently routed queries. Using these, ZeroRAG steers each query through a best-first walk that progressively rewires the overlay toward semantically relevant neighbors, so recall improves over time as the network self-organizes. We evaluate ZeroRAG using three real-world workloads and two decentralized baselines. We find that ZeroRAG outperforms baselines in terms of recall and required network messages, approaches the recall and LLM answer accuracy of a centralized document store, self-heals after catastrophic peer failures, and scales to 100 000 peers.