GitHub - nikvdp/searxng-ai-kit: Standalone SearXNG CLI for coding agents
Standalone SearXNG CLI for coding agents. Contribute to nikvdp/searxng-ai-kit development by creating an account on GitHub.
Self-Hosted AI Tool
Summary
Vane is an open-source, self-hosted AI answering engine that combines local LLMs (Ollama) and cloud providers (OpenAI, Claude, Gemini, Groq) with SearxNG-powered web search. It delivers cited answers with privacy, smart search modes, file uploads, and Docker deployment.
Highlights
- Supports local LLMs via Ollama and cloud providers including OpenAI, Claude, Gemini, and Groq with mix-and-match model selection
- Privacy-focused web search powered by SearxNG with support for web, discussion, and academic sources
- Smart search modes (Speed, Balanced, Quality) let users trade off between response time and answer depth
- Docker deployment bundled with SearxNG, configurable via web UI at localhost:3000
- File uploads, image/video search, domain-specific search, widgets, and local search history
auto-generated
Context
Audience
Developers and privacy-conscious users who want a self-hosted alternative to AI search engines like Perplexity, with control over which models and search sources are used
DomainAI Search and Answering
Formatopen-source software project
Accessfree, open source, self-hosted via Docker or source
Source Overview
Vane is a self-hosted, privacy-focused AI answering engine that runs on your own hardware and integrates local LLMs (via Ollama) and major cloud providers like OpenAI, Claude, Gemini, and Groq.
It offers web, discussion, and academic search with SearxNG, supports file uploads, image/video search, smart modes (speed/balanced/quality), widgets, search history, and domain-specific search. Vane is open source, installable primarily via Docker (bundled with SearxNG) or from source/non-Docker setups, with configuration done through a web UI at localhost:3000.
Topics
Related
OllamaSearxNGPerplexity AILocal LLM deploymentDocker Compose
