Took the sign to the streets of SF to talk to our customers

The docs of @typesense is so well written! love it

Shout out to @typesense for building a great product. Our engineering team loved Elastic Search and wanted to use Elastic for some of search scenarios, but since we are early and don't have a ton of users using Elastic didn't really feel like the right amount of engineering effort. Typesense is a great alternative with some good vector capabilities and runs great with our agents. Will share more around our experience with it in future.

Choosing an npm package can require comparing maintenance, bundle size, downloads, compatibility, and alternatives across many options. packrun.dev is an npm registry for AI agents that provides automated package comparisons, real-time scoring, and an MCP server for AI coding assistants. It scores packages from 0 to 100 using weighted downloads, bundle size, freshness, community, and quality factors. Key features: • Automated comparisons across 50+ categories using formula-based scoring. • Live package metrics for downloads, bundle size, and maintenance. • Alternative discovery for finding replacements for any package. • MCP tools for package search, metadata, health checks, comparisons, alternatives, category lookups, and install compatibility validation. • Sub-50ms search powered by Typesense Cloud. This project is licensed under MIT. Link in the reply 👇

This email was the highlight of my week! Font Awesome just switched from Algolia to a self-hosted @typesense cluster to power their search bar! For all the value I've gotten from @FontAwesome over the years (thank you for the icons through the bootstrap integration), I'm thrilled that I'm finally able to return the favor through the magic of open source software. @RobMadole, Principal Software Engineer at Font Awesome had this to say: "We compared Algolia, Postgres Full-Text Search, and Meilisearch to Typesense. We chose Typesense for the Algolia Vue component compatibility, the embedded vector, and the ease of deploying (snapshot to the rescue) and ability to tune results". The moment they switched over to Typesense, Rob sent me this email:

Active Search gives #Rails a single, Active Record-style interface for searching across Elasticsearch, Meilisearch, Typesense, and database full-text search. @37signals built it to let Basecamp and Fizzy swap search engines between SaaS and self-hosted versions without rewriting application code, and now Donal McBreen is coming to #RailsWorld to show you how it works. #ActiveSearch rubyonrails.org/world/2026/s…

🚨 SCALING YOUR APP'S SEARCH SHOULDN'T BE THE SCARIEST PART OF YOUR TECH STACK. I’ve been there. Fast search is great, until scaling brings ops bloat and unpredictable pricing. This is where @typesense FLIPS THE SCRIPT 🔥 Open-source and self-hostable 🧵↓ It’s written in C++ and ships as a single binary. It delivers the premium search experience your users expect, without the massive ops burden or SaaS tax. Why devs are making the switch: ➔ Open-source freedom: Host it yourself or use their managed cloud. ➔ Premium speed: Lightning-fast, instant search across your core data. ➔ Native typo tolerance: Handles misspelled queries beautifully out of the box. ➔ No ops bloat: Easier to use than Elasticsearch, with zero JVM dependencies. But the standout feature right now is their Natural Language Search. Instead of forcing users to rely on rigid keywords, Typesense lets you connect an LLM that automatically translates conversational phrasing—like: "cheap red shoes under $50 near me" ...directly into structured filters and search terms. It’s a massive win for user experience, and incredibly elegant to implement. Don't compromise on your search stack. The best part? 100% free and open-source. repo link in the 🧵↓

github.com/ekrist1/typesense… A fast, keyboard-driven terminal UI for administering a Typesense instance — built with Bubble Tea. @typesense