Skip to main content
Perplexica uses SearxNG as its search engine to retrieve web results while maintaining privacy. SearxNG is a privacy-respecting metasearch engine that aggregates results from multiple search engines.

​
What is SearxNG?

SearxNG is a free, privacy-focused metasearch engine that:
  • Aggregates results from multiple search engines (Google, Bing, DuckDuckGo, etc.)
  • Does not track or profile users
  • Provides JSON API for programmatic access
  • Supports specialized search engines including Wolfram Alpha

​
Default configuration

If you’re using the standard Perplexica Docker image, SearxNG is included and pre-configured. No additional setup is required.
When you run Perplexica with Docker: 
docker run -d -p 3000:3000 -v perplexica-data:/home/perplexica/data \
  --name perplexica itzcrazykns1337/perplexica:latest
The image includes:
  • SearxNG running on port 8080 (internal)
  • Pre-configured settings with JSON format enabled
  • Wolfram Alpha search engine enabled
  • Automatic startup with Perplexica

​
Using your own SearxNG instance

If you already have a SearxNG instance running, use the slim Perplexica image: 
docker run -d -p 3000:3000 \
  -e SEARXNG_API_URL=http://your-searxng-url:8080 \
  -v perplexica-data:/home/perplexica/data \
  --name perplexica \
  itzcrazykns1337/perplexica:slim-latest
​
SEARXNG_API_URL
string
required
The URL of your SearxNG instanceExample: http://localhost:8080

​
Requirements for external SearxNG

Your SearxNG instance must be properly configured for Perplexica to work correctly.
Ensure your SearxNG instance has:
1

JSON format enabled

Perplexica requires JSON API access. In your settings.yml:
search:
  formats:
    - html
    - json
2

Wolfram Alpha enabled

Enable Wolfram Alpha for computational queries:
engines:
  - name: wolframalpha
    disabled: false
3

Network accessibility

Ensure SearxNG is accessible from Perplexica:
  • If both run in Docker, use Docker networking
  • If Perplexica is in Docker and SearxNG is on host:
    • Windows/Mac: http://host.docker.internal:8080
    • Linux: http://<host-ip>:8080

​
SearxNG configuration

The bundled SearxNG instance uses the following configuration:

​
settings.yml

use_default_settings: true

general:
  instance_name: 'searxng'

search:
  autocomplete: 'google'
  formats:
    - html
    - json

server:
  secret_key: '<generated-secret>'

engines:
  - name: wolframalpha
    disabled: false
  • use_default_settings: Inherits SearxNG’s default configuration
  • autocomplete: Uses Google for search suggestions
  • formats: Enables both HTML and JSON output
  • wolframalpha: Required for computational queries and calculations

​
Advanced configuration

If you’re running your own SearxNG instance, you can customize:
​
search.autocomplete
string
Autocomplete providerOptions: google, duckduckgo, wikipedia, dbpedia, qwantDefault: google
​
search.formats
array
Supported output formatsRequired: Must include json for PerplexicaDefault: ["html", "json"]
​
engines
array
Enabled search enginesRequired: wolframalpha must be enabledYou can add more engines from the SearxNG defaults

​
Setting up standalone SearxNG

To run SearxNG separately from Perplexica:

​
Using Docker

1

Create settings.yml

Create a SearxNG configuration file:
use_default_settings: true

general:
  instance_name: 'searxng'

search:
  autocomplete: 'google'
  formats:
    - html
    - json

server:
  secret_key: '<generate-a-secret-key>'
  bind_address: '0.0.0.0'
  port: 8080

engines:
  - name: wolframalpha
    disabled: false
2

Run SearxNG container

docker run -d \
  -p 8080:8080 \
  -v $(pwd)/settings.yml:/etc/searxng/settings.yml \
  --name searxng \
  searxng/searxng:latest
3

Verify JSON format

Test that JSON API is working:
curl "http://localhost:8080/search?q=test&format=json"
You should receive JSON-formatted search results.
4

Configure Perplexica

Set the SearxNG URL in Perplexica settings or via environment variable:
SEARXNG_API_URL=http://localhost:8080

​
Manual installation

For manual installation, refer to the official SearxNG documentation.
Basic steps:
  1. Install SearxNG following the official guide
  2. Configure settings.yml with JSON format and Wolfram Alpha enabled
  3. Start SearxNG service
  4. Configure Perplexica to use your SearxNG URL

​
Configuring in Perplexica

You can set the SearxNG URL through:

​
Settings UI

  1. Navigate to Settings in Perplexica
  2. Find the “Search” section
  3. Enter your SearxNG URL in the “SearXNG URL” field
  4. Save changes

​
Environment variable

Set SEARXNG_API_URL before starting Perplexica: 
export SEARXNG_API_URL=http://localhost:8080
Or with Docker: 
docker run -d -p 3000:3000 \
  -e SEARXNG_API_URL=http://localhost:8080 \
  -v perplexica-data:/home/perplexica/data \
  --name perplexica \
  itzcrazykns1337/perplexica:slim-latest

​
Troubleshooting

​
SearxNG not responding

Test the SearxNG endpoint:
curl http://localhost:8080
You should see the SearxNG web interface HTML.
Ensure JSON format is enabled:
curl "http://localhost:8080/search?q=test&format=json"
If you get an error about unsupported format, check your settings.yml.
From inside the Perplexica container:
docker exec -it perplexica curl http://localhost:8080
If this fails, verify the URL configuration and Docker networking.

​
Wolfram Alpha not working

Wolfram Alpha is required for computational queries and calculations. Without it, some search features may not work correctly.
Verify Wolfram Alpha is enabled in settings.yml: 
engines:
  - name: wolframalpha
    disabled: false

​
Connection refused errors

If Perplexica can’t reach SearxNG:
  • Both in Docker: Use Docker network or container names
  • SearxNG on host: Use host.docker.internal (Windows/Mac) or host IP (Linux)
  • Different machines: Ensure firewall allows connections

​
Performance tuning

For better performance with your own SearxNG instance:
  • Limit the number of search engines to reduce latency
  • Enable caching in SearxNG settings
  • Use a local Redis instance for SearxNG caching
  • Configure rate limiting to prevent overload

​
Next steps

Build docs developers (and LLMs) love

Get started for freeTalk to us