How divepool works

Divepool helps you discover others, and yourself: accounts, posts, podcast episodes, and Hacker News discussions that match what you care about, found by semantic search across the AT Protocol network and beyond. This page is how it works.

How it works

We continuously pull posts — and other public records, like long-form articles and profiles — from across the AT Protocol network, the open social system Bluesky runs on, plus podcast RSS feeds and Hacker News. Everything long enough to carry meaning (at least 120 characters, in any language) gets embedded: each text becomes a numeric vector that captures what it’s about, so that items on similar topics end up close together in vector space — whether they’re a Bluesky post, a blog article, a podcast episode, or a Hacker News thread.

For each account, we group its texts into clusters based on what they’re about. Accounts with fewer than 20 embedded texts skip clustering. Every cluster then gets two summaries: a medoid — the real post closest to the cluster’s centre — and a short list of keywords, pulled out by comparing word use inside the cluster against the rest of the corpus. The medoid is the representative post; the keywords are the topic label.

All the surfaces are built on the same machinery. Type a topic and we turn your query into a vector, find the nearest items across all sources, then cluster and label them — so you get themed groups, not a flat list. Type a handle instead and you see that account’s own clusters with their keyword labels. And the personalised feed turns your clusters into a stream of new items on the same themes, with accounts you block or mute filtered out alongside items we’ve already shown you.

Try it

Type a Bluesky handle to see what an account consistently posts about, or a topic to find posts, episodes, and discussions on that subject across the network. Either way you get themed clusters with keyword labels. There’s also a discovery feed on Bluesky — the public, pre-baked version — and its signed-in twin at /feed, which isn’t limited to posts.

For agents and developers, divepool speaks MCP — connect to https://divepool.com/mcp, or point your agent at /llms.txt.

Caveats

• Text similarity, not engagement. We rank on what posts say, not on likes or reposts.
• Recency is a tiebreaker, not a driver. Fresh items get a gentle nudge at most — a great post from last week can beat a mediocre post from an hour ago.
• Coverage is partial. Not every account is synced, and short posts don’t make it past the 120-character floor.

About

Divepool is built by Jasper Rädisch, who also makes PingArthur (AT Protocol uptime monitoring), Is Bluesky Dying (network-wide churn signal) and ibla.sh (music ratings and discovery on the AT Protocol).

The method as a talk: Keywords vs Embeddings, ATmosphereConf 2026.

Spot a problem? DMs are open: @divepool.social .