Bunny.net Wants to Be Your Whole Stack — Starting With the Database

The Macro: The SQLite-at-the-Edge Gold Rush Is Already Crowded

Sometime in the last three years, the developer tools world collectively decided that SQLite, a file-based database originally designed for embedded systems and local apps, was actually the correct answer for a surprising number of web workloads. The logic holds up: it’s fast, it’s simple, it has zero infrastructure overhead, and most applications don’t have write patterns that demand the complexity of Postgres or MySQL. The problem was always “but how do you put that on the internet, distributed, without it falling apart?”

Several companies have been working on exactly that answer. Cloudflare D1, Turso, and Fly.io’s LiteFS are the most prominent names in the SQLite-compatible-but-actually-web-native category. Turso, built on libSQL (a fork of SQLite), has made the most aggressive play for developer mindshare. Embedded replicas, per-tenant databases, the works. Cloudflare D1 benefits from sitting inside an infrastructure network that already owns a large chunk of edge compute. These are not weak competitors.

The broader developer tools market gives some sense of the tailwinds. According to Mordor Intelligence, the software development tools market is expected to grow from around $6.41 billion in 2025 to $7.44 billion in 2026. Modest but steady. The more interesting pressure is architectural: the shift toward edge computing means latency isn’t just a performance metric anymore, it’s a product decision. When your users are in São Paulo and your database is in us-east-1, someone notices.

The “why now” for Bunny specifically is less about market timing and more about product expansion logic.

Bunny.net already runs CDN, storage, edge scripting, and video delivery. Adding a database isn’t a pivot. It’s the next obvious node in the graph. The real question is whether developers will consolidate around a single platform’s version of each primitive, or keep stitching best-of-breed solutions together. That question doesn’t have a clean answer yet, and I don’t think anyone pretending otherwise is being straight with you.

The Micro: One Network, One Write Lock, and a Bet on Read-Heavy Apps

Bunny Database is a managed, globally distributed database service compatible with SQLite’s query interface but designed to run over HTTP. Accessible from anywhere, no file-system assumptions required. It spins down when idle, which keeps costs low for low-traffic projects, and lets you add geographic regions incrementally without rearchitecting your schema or your application code. That last part matters more than it sounds. One of the consistent pain points with distributed databases is that you often have to plan for global distribution from day one or face a painful migration later. Bunny sidesteps that.

The technical foundation, according to a Hacker News thread on the launch, is libSQL. The same open-source SQLite fork that Turso uses.

That’s not a knock. libSQL is solid and actively maintained. But it does mean Bunny’s differentiation isn’t at the database engine level. It’s at the infrastructure and experience level, which brings us back to their CDN roots. Bunny already operates a global network with points of presence across many regions. A database product built on top of that network has a plausible low-latency story without needing to build the physical infrastructure from scratch. That’s a real advantage, and it’s the kind that compounds quietly.

The single-writer model is the honest limitation here. Bunny’s own documentation notes the product “works best for read-heavy use cases with fewer concurrent writes.” That’s not a flaw so much as a SQLite inheritance. Single-writer is baked into how SQLite works. For blogs, marketing sites, internal tools, and read-heavy APIs, this is fine. For anything with serious write concurrency, it’s a ceiling.

The launch got solid traction on Product Hunt. The comment volume was low enough that reading meaningful community sentiment from it is hard. This looked more like an announcement to an existing user base than a cold acquisition play.

The Verdict

Bunny Database is a competent, well-positioned product from a company that has already demonstrated it can run global infrastructure reliably. The SQLite-over-HTTP pattern is proven. The edge distribution story is genuinely useful. The incremental-region model removes a real friction point that competitors have mostly ignored.

What it isn’t, at least not yet, is obviously better than Turso or Cloudflare D1 for a developer starting fresh.

The differentiation case right now is simple: you might already be using Bunny for CDN or storage, and consolidation has real value. That’s a reasonable pitch for existing customers. It’s not a reason for someone with no Bunny.net account to switch over from a setup that’s already working.

At 60 to 90 days, I’d want to know whether the developer experience holds up under real workloads. Specifically around regional replication behavior and write-conflict handling. The single-writer model is fine until it isn’t, and how gracefully a product communicates those constraints matters more than most people admit before they hit them.

A few things I’d want answered before a full endorsement: pricing relative to Turso’s free tier, which is genuinely generous, actual latency benchmarks across regions, and whether the HTTP-based access model introduces meaningful overhead for write-heavy bursts. None of those are disqualifying questions. They’re just unanswered ones. Bunny is a real company with real infrastructure building a real product. That’s a better starting position than most things that ship in a given week.