Repolode
Submit
ItzCrazyKns

ItzCrazyKns/Vane

Self-hostable AI answering engine that searches the web via SearxNG and cites sources, working with local LLMs through Ollama or cloud providers.

Open-source alternative to:Perplexity

View on GitHub
Stars
37,467
Forks
4,171
Open issues
362
Last commit
2 months ago

What is it?

Vane is a privacy-focused answering engine that runs on your own hardware. It combines web search with a large language model to produce answers that cite their sources, and it can use local models served by Ollama or hosted providers such as OpenAI, Anthropic Claude, Google Gemini and Groq.

Web results come from a bundled or external SearxNG instance, so queries are routed through multiple search engines rather than a single provider. Search modes trade off speed against depth, and you can restrict queries to web, discussion or academic sources. Uploaded PDFs, text files and images can be questioned directly, and searches are stored locally in history.

Vane can be started with a single Docker command that includes SearxNG, or run from source with Node.js. It also exposes a search API for building the same answering flow into other applications.

Key features

  • 01Works with local models through Ollama and with cloud providers such as OpenAI, Claude, Gemini and Groq.
  • 02Answers include cited sources drawn from SearxNG web results.
  • 03Speed, Balanced and Quality search modes trade latency for depth.
  • 04Searches can target web, discussions or academic papers, and specific domains.
  • 05Upload PDFs, text files and images to ask questions about them.
  • 06Search history is saved locally and an API is provided for integration.

Use cases

  • Running a private, self-hosted search assistant on a home server or workstation.
  • Asking questions about your own PDFs or documents with cited answers.
  • Pairing a local Ollama model with web search without sending queries to a single search vendor.
  • Embedding an AI search flow into another application through the provided API.

Quick start

  1. 1

    Install Docker and make sure it is running

  2. 2

    Run the Vane container with bundled SearxNG, then open http://localhost:3000 and set up your AI provider

    docker run -d -p 3000:3000 -v vane-data:/home/vane/data --name vane itzcrazykns1337/vane:latest

Things to know

  • You must configure an AI provider, either a local server such as Ollama or a cloud API key.
  • The non-Docker setup requires installing SearxNG yourself with JSON output and the Wolfram Alpha engine enabled.
  • Authentication is not yet available, as it is listed under upcoming features.

Topics

#ai-search-engine#llm#rag#self-hosted#searxng#ollama#privacy#answering-engine#ai-agents#artificial-intelligence

FAQ

What is Vane?
Self-hostable AI answering engine that searches the web via SearxNG and cites sources, working with local LLMs through Ollama or cloud providers. Vane is a privacy-focused answering engine that runs on your own hardware. It combines web search with a large language model to produce answers that cite their sources, and it can use local models served by Ollama or hosted providers such as OpenAI, Anthropic Claude, Google Gemini and Groq.
Is Vane free to use? What license does it have?
Vane is open source under the MIT license. Check the license terms before using it in commercial products.
Is Vane still maintained?
Updated within 6 months. The last commit to Vane was on Sep 1, 2026.
What can I use Vane for?
Running a private, self-hosted search assistant on a home server or workstation. Asking questions about your own PDFs or documents with cited answers. Pairing a local Ollama model with web search without sending queries to a single search vendor. Embedding an AI search flow into another application through the provided API.
How do I install Vane?
Install Docker and make sure it is running Run the Vane container with bundled SearxNG, then open http://localhost:3000 and set up your AI provider: docker run -d -p 3000:3000 -v vane-data:/home/vane/data --name vane itzcrazykns1337/vane:latest
Where can I get Vane?
The source code and installation instructions are on GitHub: https://github.com/ItzCrazyKns/Vane

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects