MevaSearch
DOCUMENTATION

MevaSearch Orbit

Orbit lets you decide how MevaSearch ranks your search results: which sites to promote, which to push down and which to hide completely. You write the rules in a plain text file and Orbit applies them to every search.

What is Orbit

Orbit is a set of your own rules that changes the order of search results. For example, you can:

  • boost sites you trust (documentation, favourite sources, scientific publishers),
  • push down sites that rarely help you,
  • hide chosen domains completely,
  • limit results to your own list of sites.

Orbit is based on Brave Goggles. It uses the open Goggles rule format created by Brave Search. A file written for Brave Goggles works in Orbit, and an Orbit file written following this guide also works in Brave Search.

You can keep several Orbits, such as “Programming”, “Local news” or “No social media”, and switch between them. Exactly one Orbit is active at a time, or none.

Quick start

  1. Write the rulesCreate a text file (e.g. my-orbit.goggle) with your rules. There is an example below.
  2. Publish the filePut it somewhere publicly reachable over https://, for example on GitHub.
  3. Add it to MevaSearchOpen Settings → Orbit, click “Add Orbit”, enter a name and the file’s URL.
  4. Turn it onSelect your Orbit in the list. From now on every search uses your rules.

Your first file

An Orbit file is plain text with one rule per line. This example boosts programming documentation, pushes one site down and hides another:

! name: Programming
! description: Documentation first, no content farms
! public: false
! author: Your name

! Documentation higher
$boost=4,site=developer.mozilla.org
$boost=3,site=docs.python.org
$boost=3,site=php.net

! Less visible
$downrank=3,site=w3schools.com

! Hidden completely
$discard,site=example-content-farm.com

Lines starting with ! are metadata and comments. They describe the file and do not affect results.

Syntax

Metadata

Optional lines at the top of the file:

Line Meaning
! name: … Name of the rule set
! description: … Short description of what the file is for
! public: true / false Whether the file is listed publicly in Brave Search (not used by MevaSearch)
! author: … Author

Actions

An action follows the $ sign and is combined with a scope using a comma.

Action Effect
$boost=N Moves matching results up. N is the strength from 1 to 10. Always write the number, e.g. $boost=1.
$downrank=N Moves matching results down. N is the strength from 1 to 10.
$discard Hides matching results completely.

Scope: domain

site=domain covers the domain and all of its subdomains, so site=wikipedia.org also matches en.wikipedia.org.

$boost=5,site=wikipedia.org
$discard,site=example.com

“Only my list” mode

A bare $discard on its own line hides everything you have not boosted. That way results come only from your own list of sites:

$discard
$boost=1,site=gov.uk
$boost=1,site=europa.eu
$boost=1,site=who.int

URL patterns (advanced)

Instead of a domain you can match part of a URL. These rules are handled by the Brave engine, so they apply to Brave results in the Web tab (see Where it works).

Pattern Meaning Example
text The URL contains this fragment /blog/$downrank=2
* Any sequence of characters /docs/*/api$boost=3
^ A separator (such as /, ?, =) or the end of the URL /amp^$discard
| Start or end of the URL |https://$boost=1

By default a pattern is matched against the page URL. You can change that with $intitle, $indescription or $incontent, for example recipe$intitle,boost=2.

Where it works

MevaSearch combines results from several sources. Brave results are ranked by Brave itself, with full Goggles support. Results from the other sources are ranked by MevaSearch, using domain-based rules.

Where Support What works
Web: Brave results Full Everything in this guide, including URL patterns, $downrank and the $intitle / $incontent options
Web: other sources (WebAtlas, Wikipedia, Mojeek, Odysee) Domains $boost=N,site=…, $discard,site=… and a bare $discard
Images Domains Same as above
Videos Domains Same as above
News None Orbit does not change news results

Domain rules work in every tab: $boost=N,site=… and $discard,site=…. If a rule must apply everywhere, write it that way.

Ready-made examples

No social media

! name: No social media
$discard,site=facebook.com
$discard,site=instagram.com
$discard,site=tiktok.com
$discard,site=x.com
$discard,site=pinterest.com

Polish sources first

! name: Poland
$boost=3,site=gov.pl
$boost=2,site=pap.pl
$boost=2,site=pl.wikipedia.org
$boost=1,site=edu.pl

Science and research

! name: Science
$boost=4,site=arxiv.org
$boost=4,site=nature.com
$boost=3,site=pubmed.ncbi.nlm.nih.gov
$boost=2,site=scholar.archive.org

Official institutions only

! name: Institutions
$discard
$boost=1,site=gov.pl
$boost=1,site=europa.eu
$boost=1,site=un.org

Hosting the file

The file must be publicly reachable at an address starting with http:// or https://, and return plain text (not an HTML page).

  • GitHub: create a repository or a Gist, add the file and use its raw URL, e.g. https://raw.githubusercontent.com/user/repo/main/my-orbit.goggle.
  • GitLab: the same, using the raw URL.
  • Your own server: any URL that returns the text file itself.

If you also want to use the same file in Brave Search, keep it on GitHub or GitLab, which are the safest choices for Brave.

After you edit the file, MevaSearch picks up the new version within 6 hours (see Technical limits).

Privacy

  • Your Orbit list stays in your browser. Names, URLs and which Orbit is active are not stored in your account.
  • The file is downloaded by the MevaSearch server, not your browser. Whoever hosts the file does not see your IP address.
  • The file URL is passed to search engines. For Web searches, the URL of the active file is sent to Brave Search, which applies the rules itself.
  • The file is public. Anyone who knows the URL can read it, so do not put anything private in it.

Technical limits

What Limit
File URL http:// or https://, up to 2,000 characters
Download Up to 4 seconds, at most 3 redirects
Refresh MevaSearch keeps a copy of the file for 6 hours, then downloads it again
File unavailable The last saved copy is used. If there is none, results are shown without Orbit.
Rule strength $boost and $downrank: from 1 to 10

Troubleshooting

Orbit doesn’t change my results

Check that the Orbit is selected as active in Settings → Orbit. Open the file URL in your browser: you should see only the text of the rules, not an HTML page. On GitHub, use the “Raw” button.

I edited the file but results are the same

MevaSearch refreshes its copy of the file every 6 hours. To see changes right away, add any parameter to the URL, such as ?v=2, and save the Orbit again.

A rule works in Web but not in Images or Videos

Images and Videos only support domain rules. Rewrite the pattern as $boost=N,site=domain or $discard,site=domain.

With Orbit on, there are almost no results

Your file probably contains a bare $discard line. That is the “only my list” mode, which hides everything except boosted domains. Add more domains with $boost or remove that line.

Can I use Goggles made for Brave Search?

Yes. Orbit is based on the Brave Goggles format, so just add the Goggle’s file URL in Settings → Orbit. It works most completely in the Web tab.