I don’t think WOT will work very well for a short form video app with a mostly asymmetric follow graph. But we’ll take a look. It works much better with smaller networks with mutual trust I think.
Login to reply
Replies (2)
WoT has multiple uses. The first is getting rid of the obvious bots and bad actors who may want to spam your app with junk. Centrality algorithms like GrapeRank or PageRank have a much more extensive reach than follows of follows, and are effective and simple enough to implement if and when spam becomes a problem on your app. Ideally you’ll farm out calculation of personalized trust metrics to a WoT Service Provider like Brainstorm, Vertex, or Relatr. @Amethyst and other clients have already started doing this using @Vitor Pamplona's Trusted Assertions NIP (which @fiatjaf just merged) to empower the user and avoid SP vendor lock-in. Cultivation of the ecosystem of WoT SPs to make WoT metrics both *personalized* and *portable* is the primary goal of the #wotathon — the @nosfabrica web of trust hackathon that runs through April. We have weekly community calls on Thursdays; you should join us!
The more challenging and exciting WoT use case will be content discovery. Suppose I want to find songs in some niche musical category by awesome but undiscovered musicians who aren’t on Spotify. Do we want to wait for said musicians to build up 10k followers on nostr? Why wait for that when my trusted, extended community can find exactly the content I want and serve it up to me on a silver platter? This is the type of use case we’re tackling with Brainstorm at @nosfabrica. This particular use case is being explored in depth by @Avi Burra. Keep an eye on what he’s up to; it will be relevant to your app. 2026 will be an exciting year for WoT in nostr!
cc @Jon Gordon


NosFabrica - Weaving the Fabric of Trust on Nostr
Wotathon - NosFabrica
November 2025 – April 2026 TOTAL REWARD POOL: 50 MILLION SATOSHIS GOAL: Develop the Web of Trust ecosystem for Nostr, build open source libraries...
Depends the use case. For example for recommendations it's going to work quite well imo