tl; dr: Spyglass is a search platform that lives on your device, indexing what you want, exposing it to you in a super simple & fast interface.

⚠️ Spyglass is very much in its early stages, but it’s in a place where it's functional and can be used to replace basic searches. ⚠️

Stable compiled builds are provided on the releases pages. Download the appriopriate file for your OS (e.g. .deb for linux, .dmg for macOS and .msi for Windows)

If you're interested in building from source, after checking out the repository run the following:

make setup-dev
make build-release

Once launched, press Cmd + Shift + / to open Spyglass. Queries prefixed with / will search through your installed lenses, otherwise it'll search through your index.

Use the arrow keys to select the result you want and hit Enter to open the link in the browser of your choice!

Spyglass is a solution to address the following common issues when searching the web.

• Do you add terms such as reddit or wiki to your searches to narrow it down?
• Do you skip over a full-page of ads before getting to your actual search results
• Do you scroll past dozens of SEO spam pages to find the recipe/review/blog post you were looking for?
• Do you get frustrated with overzealous autocorrect on your search terms?

Spyglass expands on the ideas outlined in this paper by the Brave Search Team.

You can add different lenses that clue the application into what you want to have indexed. Here are some examples that I've been personally using:

Interested in cooking & recipes? Add a "recipe" lens which will go index a curated set of websites with high quality recipes.

(
    version: "1",
    name: "recipes",
    description: Some(r#"
        A curated collection of websites with useful, high-quality recipes.
    "#),
    domains: [

        # Major sites that often have really good recipes
        "www.seriouseats.com",
        "cooking.nytimes.com",
        ...

        # Specific cuisines/sites that I've found randomly w/ high-quality recipes
        "www.hungryhuy.com",
        "www.vickypham.com",
    ],

    # Not yet supported but ideally more ways to filter URLs within a domain
    urls: [
        "www.reddit.com/r/recipes/*",
    ]
)

Interested in the Rust programming language? Add the "rustlang" lens which will index the Rust book, rust docs, crate.io, and other sites that are related to the programming language and not the Rust game / The Rust Belt / oxidation / etc.

(
    version: "1",
    name: "rustlang",
    description: Some("Rustlang targeted websites"),
    domains: [
        # Support for wildcards in domain names
        "*.rust-lang.org",
        "docs.rs",
        "rustconf.com",
        "crates.io",
        "this-week-in-rust.org",
        ...
    ],

    # Again not yet supported but an example of indexing specific communities that
    # are relevant to the topic
    urls: [
        "www.reddit.com/r/rust",
        "www.reddit.com/r/rust_gamedev",
        "https://github.com/topics/rust"
    ]
)

The settings.ron file can be found by "Show Settings folder". If there is no file found in their directory on startup, a default one will be created.

(
    # The max number of pages to index per domain
    domain_crawl_limit: Finite(1000),
    # The max number of crawlers per domain
    inflight_domain_limit: Finite(2),
    # The max number of crawlers in total
    inflight_crawl_limit: Finite(10),
    # Not used... yet!
    run_wizard: false,
    # Not used... yet!
    allow_list: [],
    # Domains to completely ignore.
    block_list: [
      "web.archive.org",
      "w3schools.com"
    ],
)