Bluesky Spatial Semantic Search

I’ve combined my 1) desire to re-surface content from my Bluesky feed 2) embeddings & semantic search and 3) my love of infinite canvas UIs:

It’s an OpenAI embeddings-powered search tool for my Bluesky likes, with advanced filtering, on an infinite canvas with draggable results. I started this project last spring for my Twitter likes + bookmarks, and since adapted it to Bluesky.

Right now you can only run it locally or self-host it; I’m not offering a hosted version.

Running it yourself

(Using Bun because it’s excellent for scripting)

1. Clone repo, cd
2. bun i
3. Copy .env.sample into .env, and fill out the API credentials. You’ll need an OpenAI API key, your Bsky handle (e.g. lachlanjc.com), plus a Bsky App Password
4. Install Chroma, then in another terminal, chroma run (you’ll need to keep this running both for the download scripts and querying the embeddings database)
5. If you’re using Zed, spawn the task for download, then for embedding. Otherwise, bun run lib/tasks/download.js then bun run lib/tasks/embed.js
6. bun dev & open the link!

Open source libraries

• Next.js site
• Chroma for embeddings database
• umap-js for positioning posts
• React Flow for infinite canvas
• bsky-react-post for post embeds
• Radix Themes for search UI

---

MIT License