December 18, 2025

Lost in translation, found in comments

Fluent: A Localization System for Natural-Sounding Translations

Fluent promises natural translations — devs flex, AI stans argue, users want tools

TLDR: Fluent lets apps speak naturally across languages with smart rules for plurals and gender. The thread erupted into an AI-versus-handcrafted debate, while real users complained about missing tools to validate and edit files—making the tech promising, but the tooling gap the showstopper.

Meet Fluent, the open‑source system trying to make apps sound like real humans in any language. It lets translators write naturally (plural forms, gender, style) without being chained to English, and supports friendly syntax with JavaScript, Python, and Rust. Sounds wholesome… until the comments lit up.

First up, a classic dev flex: one commenter waved the flag for their 2017 project, Lokalized, saying it had “the same goal” but a different approach. Cue the crowd: is Fluent reinventing the wheel, or finally doing it right? Then the AI vs. rules fight broke out. An AI fan dropped the mic with “tiny LLMs” (small large language models) could do this better, sparking a slap-fight over reliability. Team Deterministic shouted “we need guarantees,” while Team AI teased, “let the bots conjugate.”

Meanwhile, the people actually shipping products showed up with receipts: “I use this,” said one dev, “but the tooling is weak.” No file validators, no editor support, and POedit (a popular translation app) basically ghosted the format. The memes rolled in: “One does not simply ship i18n without tools,” and “tabs need therapy” after Fluent’s plural examples. Verdict from the crowd: Fluent’s idea is loved, but the tooling drama is the real plot twist.

Key Points

  • Fluent enables asymmetric translations so target languages can use genders, cases, and variants independent of the source language.
  • Progressive enhancement isolates locale logic, letting translators add complexity per language without affecting others.
  • The system supports date/time/number formatting, plural categories (via CLDR), bidirectionality, custom formatters, and robust runtime handling.
  • Fluent Syntax 1.0 was released in April 2019; implementations exist in JavaScript, Python, Rust, with React bindings, under Apache 2.0.
  • Examples show message identifiers, variable-driven plural variants, and terms for consistent reuse and branding in applications like Firefox.

Hottest takes

“has the same goal but took a different approach” — revetkn
“This problem is more suitable for tiny LLM.” — mlajtos
“it lacks any tooling around it” — skwee357
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.
Fluent: A Localization System for Natural-Sounding Translations - Weaving News | Weaving News