vane-search

Query a local Vane service for citation-backed answers with source summaries.

103|14|Updated Feb 22, 2026
One-click install
npx skills add https://github.com/szsip239/teamclaw --skill vane-search
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: vane-search
Source: https://github.com/szsip239/teamclaw/tree/main/data/skills/vane-search
Command: npx skills add https://github.com/szsip239/teamclaw --skill vane-search

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Vane Search enables local, citation-backed information retrieval by querying a locally deployed Vane service and returning results with sources and summaries.

Core Features & Use Cases

  • Local deployment with a fast, configurable search backend (SearxNG) combined with embedding-based reordering and LLM summarization to produce citations.
  • Supports multiple sources (web, academic, Reddit) and three depth modes (speed, balanced, quality) for flexible research workflows.
  • Use case examples include: conducting literature reviews, competitive research, and structured information gathering with cited sources.

Quick Start

Query the local vane API with a question and return a cited summary from the configured sources.

Frequently Asked Questions about vane-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get citation-backed search results from a locally deployed LLM service?▼

Citation-backed search results are generated by querying a local Vane service, which uses SearxNG, embedding-based reordering, and LLM summarization to return answers with source references.

Can I use local search for academic literature reviews and Reddit information gathering?▼

Local search supports academic literature reviews and Reddit information gathering by querying web, academic, and Reddit sources, returning LLM-generated summaries with cited references.

What is the best way to control search depth for complex research questions?▼

Search depth for complex research questions is controlled using three configurable modes: speed, balanced, and quality, allowing flexible workflows for information retrieval and deep analysis.

How does embedding-based reordering work for information retrieval?▼

Embedding-based reordering in information retrieval improves result relevance by reorganizing SearxNG search outputs before LLM summarization, ensuring the most pertinent sources are cited.

Do I need a local API endpoint to run citation-backed searches?▼

A local API endpoint at http://localhost:3010/api/search is required to run citation-backed searches, with an accompanying web UI at http://localhost:3010 for direct interaction.

When should I not use a local search deployment for deep analysis?▼

Local search deployment for deep analysis is not suitable when external API access is restricted or when the hardware cannot support local LLM summarization and embedding processing workloads.