Skip to main content
Back to trending

RustXNG

Self-hosted, SSRF-hardened web search and grounding engine for local AI agents (AGPL-3.0-or-later)

  • agpl
  • ai-agents
  • grounding
  • llm
  • mcp
  • metasearch
  • privacy
  • rag
  • rust
  • search-engine
  • searxng
  • self-hosted
View on GitHub
Stars
365
Forks
22
+ today
+8
Created
15d

Ranking data as of November 1, 2026 (UTC).

Star History

Today, hour by hour

365 stars
08

Overview

Self-hosted, SSRF-hardened web search and grounding engine for local AI agents (AGPL-3.0-or-later) It ranks #45 on GitTiger, gaining +8 stars on November 1, 2026 (UTC).

The project is written in Rust and has 22 forks. It was created 15d ago.

AI-generated summary of the repository README — may contain errors.

What it is

RustXNG is a self-hosted web search and grounding engine written in Rust for local AI agents. The description says it is hardened against SSRF and carries the AGPL-3.0-or-later license; its topics mention metasearch, SearXNG, MCP, RAG and privacy.

Capabilities described

According to the README preview:

- It reimplements 226 engine types from SearXNG, with 352 settings entries of which 180 are on by default. - Results are re-ranked using term coverage, canonical-site authority, identifier matching such as RFC numbers, and deduplication that ignores thumbnails. - Reliability features include per-engine timeouts, suspension with backoff, and automatically resting engines that keep returning nothing. - An optional local index, rustxng-index, can crawl your own corpus, rank it with BM25 and merge it into results. It is described as experimental and off by default.

Getting started

No install or run commands were supplied. Follow the repository README for build and configuration steps, including how to connect it to an agent.

Evidence gaps

Details of the MCP integration and performance claims were not supplied beyond the tags.

Installation
git clone https://github.com/sahenjp/RustXNG.git
cd RustXNG
# see README for setup

Frequently Asked Questions

What is RustXNG used for?

It is a self-hosted search and grounding engine that gives local AI agents web results. The description also stresses SSRF hardening.

How does RustXNG rank search results?

The README preview says it goes beyond raw engine order by weighing term coverage, canonical-site authority and identifier matches such as RFC numbers, and by deduplicating results without relying on thumbnails.

Is the local index in RustXNG ready for production?

The README preview labels the optional rustxng-index local index experimental and off by default. It crawls your own corpus, ranks with BM25 and merges into results.

What license does RustXNG use?

The supplied metadata lists AGPL-3.0, and the description says AGPL-3.0-or-later. Confirm the exact terms in the repository before use.