Most text search solutions are fine tuned to serve results for keyword style queries. This is a problem in the age of MCP, as comparably long user input hits the same old pg_trgm search.
In this post, we dig into why classic full-text fail and walk through a pragmatic recipe that fuses ts_rank, pg_trgm, and pgvector to restore relevance.
If you're shipping search in 2025, you'll want these fixes in place before your MCP server returns empty sets to to the waiting world.
Lucas Sifoni shows how he uses Elixir's hot code loading to modularize his monolithic app. He deploys one instance per client and dynamically loads client-specific LiveView apps at runtime without stopping the system.
Chris Gregori shares his process for going from idea to live SaaS waitlist in 2 hours. He uses AI for validation, Cursor with PRDs, Elixir/Phoenix, and his Bloom UI library for rapid prototyping and market testing.
Chris McCord introduces Phoenix.new, a browser-based coding agent designed specifically for Elixir and Phoenix development. It offers a complete virtual machine environment with root access, headless browser testing, and can build full applications while you watch (or go grab a coffee).
How well do you know Elixir's built-in telemetry and monitoring features? Phoenix gives you a complete dashboard, VM metrics, and distributed telemetry out of the box—no setup required. Learn how to use it to monitor your production apps.
Although this is not an article, the demo is so cool that I decided to feature it here. Dima Mikielewicz built this to showcase Torus, a lib that integrates PostgreSQL's search into Ecto queries. Test different search methods — semantic, full-text, and similarity — and see how each handles the same query.
Angelika Tyborska shows how to work with XML documents in Elixir using the Erlang xmerl library. Beyond XML processing, it's a great guide for using Erlang libraries in Elixir, teaching how to work with Records, for example.
Ash AI was recently launched. It makes it easy to add LLM interactions to an application, and can even expose existing secure actions to an agent acting on the user's behalf. Christian Alexander shows a demo of using Ash AI.
Our team runs a multi-tenant Phoenix app powering 3 motorcycle/powersports-centric eCommerce brands: RevZilla, Cycle Gear, and J&P Cycles. We are looking for a seasoned frontend pro to join us in delivering quality software in service of our mission to advance the experience of powersports enthusiasts across the globe.
