Writing · 2026 · 3 min

Announcing links.json: Human-Curated Links for a Community-Driven Search Engine

I want to make the web open again. When the web first started in the 1990s, it was common getting lost down a rabbit hole of random topics, from literature, fan pages and to personal sites. It was exciting opening a search engine and just not knowing what you would find. There was no algorithm dictating what is allowed to be shared and who sees it. This has slowly decayed to the point where many social media sites, like Facebook, discourage link sharing via algorithm degrading your posts – because they want you to stay on their site – but more recently are even now charging for sharing links in groups.

Take a second to think about the last time Youtube, Facebook, Instagram, Reddit or an other social media has actually sent you to a cool independent website? Not a business or a news article. An actual website built by a person sharing their interests or opinions? If you’re in tech, maybe you read tech blogs but the point is, the ability to share links to external websites is dead. Even way back posting on any forum would allow you to share your links, because you were allowed to have a signature and so every post you made also included your links. So by contributing to a websites forum and supporting the community, you were granted the ability to share your links. Reddit doesn’t allow this, instead you could get banned for “self-promotion”. On top of that, they won’t even allow you to just post a link to your blog, you have to copy your blog post to the Reddit post. The idea was always if you contributed to a community, you got something back for it but not anymore. This madness has to end.

In 2024, I wanted to find content I enjoyed – without the algorithm. So I started a video scraping crawler, but little did I know, the internet has grown to such a size it would take a database of billions of records to properly index everything. One person or team cannot sufficiently index the net for others to search.

After all, this is the internet, it was designed to be decentralized, so I have now created a decentralized solution. Any website can create a links.json file, it contains groups of links that can be indexed by any client and it can also contain links to other links.json. This in effect creates a web of links maintained by many people, when pulled together can then essentially create a long-living dynamic index of links. Like it’s own search engine but you control it, by choosing the tags you want to include / exclude, keywords etc.

The project is called links.json and you can view it here: https://codeberg.org/bendangelo/links-json.

I already have my links at /links and at /links.json. There is currently no client but it is being worked on now.

The ultimate goal is to create a standard /links.json on blogs, independent websites and anywhere else. They can then be indexed by any user and then searched. This effectively by passes traditional search engines and with a steady stream of new links added, it can even be turned into an algorithm, but one which you control.

Shouted out to humanns.json

There is an other project with similar goals called human.json, they want to label websites as human created to avoid ai generated slop. They created a solution by having an other website vouch for each other. This in effect creates a web of human verified websites, what if these sites also have their own links.json? This would create an indexable list of links. This can be a cool experiment to jump start this project.

I’ll love to here your thoughts.

Comments

No comments yet. Be the first.

Leave a comment

Plain text. Comments are reviewed before they appear; your email is never asked for.