March 14, 2026
Database or drama base?
Postgres with Builtin File Systems
Red‑hot Postgres remix promises ‘no config, no buckets, no problem’ as devs cheer, squint and side‑eye
TLDR: A new service promises a super‑charged database that also stores files and powers AI features so developers can skip extra tools and setup. Commenters are split between calling it genius, questioning whether the marketing is misleading, and laughing that even the homepage carousel is broken on mobile.
A wild new tool is trying to turn the humble database into a one‑stop shop for everything your AI apps need: data, files, and even talking to other websites. The pitch is bold: toss your files in, stash your structured data, auto‑generate “smart” text math (embeddings), and forget about fiddling with cloud storage or extra services. One command and you’re in, they say.
But the community isn’t just clapping — it’s squinting. One user dramatically accuses the project of burying the lede, pointing out that under the shiny “Postgres” branding, the data is actually sitting on a different system entirely. Translation: “Are we being sold soda and secretly getting sparkling water?” Another commenter is fascinated that this whole thing was apparently built by a single person mostly by talking to artificial intelligence, calling out the behind‑the‑scenes story as the real headline.
Others are way more chill. Some love the idea of mixing files and database data in one place so they don’t have to juggle folders, servers, and storage dashboards anymore. A newcomer to Postgres gushes about how flexible it feels compared to their old databases. And of course, classic internet energy: while people argue about architecture, someone else shows up just to complain that the fancy image slider on the homepage is broken on their phone. Peak tech drama.
Key Points
- •Serverless PostgreSQL is provided with built-in file storage, embeddings, vector search, and environment tooling.
- •Structured state remains in Postgres while unstructured data (e.g., transcripts, artifacts) is stored as files in the same workspace.
- •Embeddings can be generated inline via an embedding() SQL function; similarity search is supported using vector operations.
- •Outbound HTTP requests can be made directly from SQL using a built-in http_get() function.
- •Setup is streamlined: one command to install, one to create a database, and no S3 configuration required for file handling.