Practical guide to configure SearXNG like Kagi (step-by-step with Docker and settings.yml)

Traditional search engines like Google and Bing have become synonymous with intrusive ads, user profiling, and limited control over search result quality. In response, new alternatives are emerging: Kagi, a paid search engine praised for its clean results and personalization, and SearXNG, a free, open-source metasearch engine focused on privacy, customization, and transparency.

SearXNG is free, but it requires some configuration to deliver high-quality results. The good news: with a few tweaks, you can adjust SearXNG to replicate much of the Kagi experience—without paying for a subscription. In this step-by-step tutorial, we’ll walk through how to deploy it with Docker, and then fine-tune your settings.yml file to filter sources, boost trustworthy sites, and block low-quality content.

---

Why Configure SearXNG Like Kagi?

Kagi has become popular thanks to its Leaderboard feature (see Kagi Leaderboard), which allows users to promote reliable websites like Reddit, Wikipedia, or Stack Overflow, while demoting content farms, spam, or irrelevant sources. This dramatically improves search quality.

While SearXNG doesn’t natively provide a Leaderboard, its flexible settings.yml configuration lets you achieve a similar effect:

• Boost specific sites to push them higher in the SERPs.
• Block unwanted domains like TikTok or spammy content mills.
• Mix engines (Google, Bing, DuckDuckGo) while avoiding reliance on a single provider.

In short: with the right setup, a self-hosted SearXNG instance can become a viable Kagi alternative.

---

Step 1 – Install SearXNG with Docker

One of SearXNG’s strengths is how easy it is to deploy with Docker. Whether you’re on a VPS, a local server, or even your workstation, here’s a minimal configuration:

version: '3'

services:
  searxng:
    image: searxng/searxng:latest
    container_name: searxng
    ports:
      - "8080:8080"
    volumes:
      - ./settings.yml:/etc/searxng/settings.yml
    restart: unless-stopped

Start the container:

docker-compose up -d

Access your local search engine:
👉 http://localhost:8080

At this stage, SearXNG is running, but the results are still “raw.” The real power lies in the settings.yml file.

---

Step 2 – Understanding the settings.yml File

The settings.yml file is the core of SearXNG customization. Here you can define:

• Which search engines to query (Google, Bing, DuckDuckGo, Wikipedia, YouTube…).
• Which sites to boost or block.
• Your language, region, and privacy preferences.

Minimal example:

general:
  instance_name: "My SearXNG"
  debug: false

search:
  safe_search: 1
  autocomplete: duckduckgo

engines:
  - name: google
    engine: google
    disabled: false
  - name: bing
    engine: bing
    disabled: false
  - name: duckduckgo
    engine: duckduckgo
    disabled: false

⚠️ Note: Some engines (especially Google) may block queries if you send too much traffic without an API key.

---

Step 3 – Recreate Kagi’s “Leaderboard” in SearXNG

Here’s where things get interesting: by adjusting site priorities, you can mimic Kagi’s Leaderboard.

Example: boost Reddit and Stack Overflow, block TikTok:

result_filters:
  - name: boost_reddit
    action: boost
    pattern: reddit.com
    weight: 2.0
  - name: boost_stackoverflow
    action: boost
    pattern: stackoverflow.com
    weight: 2.0
  - name: block_tiktok
    action: block
    pattern: tiktok.com

With these rules in settings.yml:

• You’ll see more relevant results from Reddit and Stack Overflow at the top.
• TikTok results will be fully filtered out.

On Reddit, some users have shared similar configurations, reporting that they reached Kagi-like quality after customizing SearXNG filters.

---

Step 4 – Further Customize Your Results

SearXNG offers even more ways to refine your search:

• Limit enabled engines (disable noisy or redundant ones).
• Adjust engine weights (e.g., give more importance to Bing or DuckDuckGo, less to Yahoo).
• Configure SafeSearch for sensitive filtering.
• Enable a temporary cache for faster responses.

Example: engine weighting

engine_weights:
  google: 1.5
  bing: 1.0
  duckduckgo: 0.8

By balancing engine weights and fine-tuning filters, your open-source SearXNG instance becomes a powerful, privacy-respecting search engine—one that rivals premium options like Kagi.

---

Your comments enrich our articles, so don’t hesitate to share your thoughts! Sharing on social media helps us a lot. Thank you for your support!