Finding content on the fediverse has always been one of its most stubborn practical problems. The architecture that makes the fediverse distributed and community-run also makes search genuinely hard - not just as an engineering challenge, but as a question of values. Fediscovery and related efforts are trying to build something that actually works without trampling the norms that make the fediverse worth being on. This article covers how fediverse search works today, what Fediscovery is doing differently, and what you can do right now to make your content easier to find.
Why Search Works Differently Here
On Twitter/X or Bluesky, a search query hits a single global index. Every public post is in there. The fediverse does not work that way. Each Mastodon instance runs its own database, and that database only contains content the instance has actually encountered through federation. If nobody on your server follows someone on a remote server, that person’s posts are invisible to your search. They might as well not exist.
This is not a bug that will eventually get patched. It is a direct consequence of how the network is built.
Mastodon also made full-text search opt-in per user, which reflects a long-standing community norm: not everything public is meant to be indexed and surfaced out of context. Hashtags ended up becoming the most reliable discovery mechanism as a result, partly by design and partly by necessity. They federate consistently and signal intent clearly, which free-text search across fragmented databases cannot reliably replicate.
What Fediscovery Actually Does
Fediscovery is a project building opt-in, privacy-respecting search infrastructure for the fediverse. The starting assumption is that search is only legitimate if users have genuinely chosen to participate in it.
The indexing model is opt-in by default. Users must explicitly allow their posts to be indexed. This is not a fine-print checkbox buried in settings - it is the foundational design choice, made because the fediverse community has historically been sceptical of anything that feels like surveillance dressed up as a feature.
Rather than building one central search engine to rule everything, the project aims for a federated search model where multiple providers can participate. That matters because a single centralised index would reproduce the same concentration-of-power problems the fediverse was built to avoid. The technical work builds on existing ActivityPub conventions and proposes extensions that any implementation can adopt, which means it is not a proprietary silo.
In practice, Fediscovery is designed to enable cross-instance search for opted-in content, people discovery across the wider network, and hashtag aggregation that goes beyond what any single instance can see on its own. There is also a goal of integrating with Mastodon’s built-in search UI, which would make the experience considerably less clunky than bouncing between external tools.
Current status: active development, gradual rollout, some instances testing integration. Effectiveness scales with adoption, so it will be patchy for a while.
Discovery Methods That Work Right Now
While Fediscovery continues to develop, there are several ways to find content and people on the fediverse today.
Hashtag follows are probably the most underused feature. Many Mastodon clients and instances let you follow a hashtag directly, creating a feed of every post using it that your instance has seen. For topic-based discovery, this is currently the most effective option available.
Mastodon instances also track trending hashtags, links, and posts. These are moderated by instance admins, so the quality varies, but they provide a curated snapshot of what is resonating across the network rather than just on your home timeline.
Relay servers are worth knowing about even if most users never configure one themselves. Relays share public content between instances, so an instance subscribed to a relay sees far more of the federated network. This improves both the federated timeline and search results in a way that is largely invisible to end users but makes a real difference.
Mastodon’s profile directory - available on many instances - lets users opt into being listed by topic and activity level. It is a simple but genuinely useful tool for finding accounts in specific areas.
For anything beyond the built-in tools, several community-built third-party applications aggregate fediverse content for discovery purposes. The tools guide covers current options.
The Privacy Trade-off
Search and discovery create tension with privacy almost by definition. A user who wants to be found benefits from comprehensive indexing. A user who values contextual integrity - the idea that posts shared in one context should not automatically travel to others - may feel differently. Both positions are reasonable.
The fediverse community has historically prioritised consent over discoverability, which is why Mastodon’s search has been conservative compared to centralised platforms. Fediscovery’s opt-in approach tries to hold both values at once by giving users explicit control rather than defaulting to maximum indexing and hoping nobody notices.
Practically speaking: enable the “discoverable” flag on your profile if you want to appear in searches and directories. Opt into search indexing if your instance and client support it. Use hashtags on posts you want others to find. And be clear in your own mind that unlisted and followers-only posts should never be indexed by any legitimate service - that is not a gap in the system, it is the system working correctly.
Making Your Content Easier to Find
Search infrastructure aside, there are things you can do today that will remain useful regardless of how discovery evolves.
Hashtags are still the primary mechanism, and that will not change even if full-text search improves dramatically. Use them on anything you want discovered, and be specific rather than generic - niche hashtags reach interested people more reliably than broad ones. Setting your profile to discoverable opts you into instance directories and relevant search features with one toggle. A detailed bio matters too: people searching for accounts in your area should find the relevant terms there, not have to infer them from your posts.
Public posts are the only ones that can be indexed or appear in federated timelines, so if you want reach, post publicly. And consistent participation in hashtag communities builds visibility over time in a way that passive posting does not. Accounts that engage tend to be better known than accounts that only broadcast.
Better discovery also has a downstream effect on your timeline experience. Finding more relevant accounts to follow improves your home timeline directly. When your instance has richer federated data - through relays or search integration - the federated timeline becomes less of a firehose and more of something you actually want to read.
Common Mistakes
The most common mistake is expecting something close to Google-quality instant results. Fediverse search is improving, but it will remain more limited than centralised search for the foreseeable future, and adjusting expectations accordingly saves a lot of frustration.
Related to that: abandoning hashtags on the assumption that full-text search should handle discovery. Hashtags are essential today, and they will complement rather than be replaced by full-text search - they do different things. Similarly, assuming all public content is searchable is wrong; only opted-in public content gets indexed by opt-in services.
Ignoring the discoverable profile flag is a surprisingly common oversight. It takes seconds to enable and meaningfully improves your visibility in directories and search features. Finally, it is worth being clear that opt-in search and surveillance are not the same thing. The consent model exists precisely to preserve the distinction.
Frequently Asked Questions
Does Fediscovery index private posts? No. Only public posts from users who have opted into search indexing are included. Private, followers-only, and unlisted posts are never indexed.
Can I opt out of Fediscovery? Indexing is opt-in by default, so there is nothing to opt out of unless you have already chosen to participate. If your instance integrates with Fediscovery, your level of participation is controlled through your account settings.
Will search replace hashtags? No. Hashtags provide explicit categorisation that complements free-text search rather than duplicating it. Both tools will remain relevant, and they work best used together rather than treated as alternatives.
How does this affect my instance’s resources? Integrating with search infrastructure adds some resource overhead. Instance admins can weigh this against their server capacity and community needs. The developer notes cover the technical side in more detail.
Is there a search API for developers? Mastodon’s API includes search endpoints. Fediscovery aims to extend this with cross-instance capabilities. Developers should check current API documentation for what is available at any given point in the rollout.
When will cross-instance search be widely available? Adoption depends on instance admin decisions and user opt-in rates. A gradual rollout is expected throughout 2026 and beyond. The articles hub will carry updates as things develop.