Top AI Product

Every day, hundreds of new AI tools launch across Product Hunt, Hacker News, and GitHub. We dig through the noise so you don't have to — surfacing only the ones worth your attention with honest, no-fluff reviews. Explore our latest picks, deep dives, and curated collections to find your next favorite AI tool.


fx by Vercel Labs — tiny native coding agent written in Zig ships as a 7.8MiB binary with 10µs cold starts

Claude Code burns ~33k tokens on startup and drags a whole Node runtime with it. fx is Vercel Labs betting on the opposite extreme: a coding agent as one native binary — 7.8MiB, 10 microsecond cold start, single-digit megabytes of memory. Originally an internal Vercel tool, now Apache-2.0 on GitHub: 1.2k stars in days, 199 points on HackerNews.

A harness, not another terminal IDE

The CLI feels like a Unix shell, not a heavyweight TUI. But the real positioning is “embeddable harness”: something you drop into agent sandboxes, CI, or resource-constrained environments where packing hundreds of agent instances per machine actually matters. Model and provider agnostic, local or cloud inference. No telemetry — sessions stay on your machine.

How you plug into it

One curl command installs it. For embedding, there’s a WebAssembly SDK (fx-core.wasm), Agent Client Protocol for editor integration, plus MCP and subagent support. Your app supplies the network transport and permissions; fx supplies the agent loop.

The honest catch: inference currently routes through Vercel AI Gateway, and HN called that lock-in loudly. More providers are promised. If they land, this becomes the SQLite of coding agents.


You Might Also Like


Discover more from Top AI Product

Subscribe to get the latest posts sent to your email.



Leave a comment