Do coding agents recommend GNU gettext?
Coding agents raised GNU gettext in 164 of 500 judged internationalization sessions and chose it alone in none of them, and only as part of a joint choice in 1.
Coding agents raised GNU gettext in 164 of 500 judged internationalization sessions and chose it alone in none of them, and only as part of a joint choice in 1.
This page reports what happened when Claude Code, Codex, Cursor, Grok Build CLI and Muse Code had to solve a problem in internationalization inside a realistic codebase. Not what a chat assistant says about GNU gettext. What an agent actually installed.
The numbers
| Category | Internationalization |
| Sessions in the category | 500 |
| Sessions where GNU gettext was chosen | 0 |
| Install share | 0% |
| Raised as a candidate, not chosen | 164 |
| Chosen when considered | 0% |
| Site | gnu.org |
Which agents raised GNU gettext
Every agent considered it and none chose it alone. Cursor raised it most often, in 48% of its internationalization sessions.
| Agent | Sessions | Raised GNU gettext | Chose it alone |
|---|---|---|---|
| Claude Code | 113 | 29 (26%) | 0 |
| Codex | 150 | 44 (29%) | 0 |
| Cursor | 103 | 49 (48%) | 0 |
| Grok Build CLI | 26 | 8 (31%) | 0 |
| Muse Code | 108 | 34 (31%) | 0 |
Which buyers it came up for
It surfaced most for requests written as enterprise team, in 64% of those sessions. Knowing which buyer already has the product in mind tells you which pages to fix first.
| Who is asking | Sessions | Raised it |
|---|---|---|
| Vibe coder | 102 | 2 (2%) |
| Junior developer | 115 | 26 (23%) |
| Senior engineer | 113 | 28 (25%) |
| Enterprise team | 170 | 108 (64%) |
What GNU gettext was up against
The full ranking in internationalization, from the same sessions:
| # | Product | Runs won | Share |
|---|---|---|---|
| 1 | next-intl | 110 | 22% |
| 2 | Django translation framework | 96 | 19% |
| 3 | Symfony Translation | 41 | 8% |
| 4 | Paraglide JS | 41 | 8% |
| 5 | Built in-house (no product adopted) | 41 | 8% |
| 6 | Rails I18n | 38 | 8% |
| 7 | Vue I18n | 27 | 5% |
| 8 | i18next | 27 | 5% |
What this means
Raised 164 times and never chosen alone is a specific, diagnosable result, and a better starting position than being unknown. The agent has the product in mind. It reached the shortlist 164 times and made it into 1 joint choice, but never into a choice of its own.
In our data the causes, in the order they occur:
- The quickstart does not run when pasted, so the agent abandoned it mid-integration.
- The documentation describes an interface that changed, so the generated code failed.
- The product name and the package name differ, so the install step went wrong.
- The fit was genuinely wrong for the repository, which is fine and worth knowing.
All but the last are fixable in days, and the reason is written down in the session transcript.
Where these numbers come from
The 500 sessions in internationalization are part of a published set of 15,051, run with real coding agents inside realistic codebases and judged blind. The full method is on one page: how we measured this.
Every internationalization run can be replayed on the board.
If you work on GNU gettext: the judge recorded a reason for every session where it was raised and passed over. Those reasons are in the transcripts.
<!-- generated by scripts/write-data-pages.mjs -->
Common questions
Do coding agents recommend GNU gettext?
They raise it but rarely choose it. Across 164 sessions where GNU gettext came up as a candidate, agents never chose it alone; it was part of a joint choice in 1.
Does Claude Code recommend GNU gettext?
In 0 of the 113 sessions in internationalization run with Claude Code, which is 0%.
Do different coding agents treat GNU gettext differently?
Not much. Claude Code, Codex, Cursor and Muse Code chose it at similar rates, between 0% and 0% of their runs.
How was this measured?
Real coding agents at pinned versions were run in sandboxes inside 92 realistic codebases and asked to solve real tasks. A simulated project owner approved or questioned each recommendation before any code was written, and a judge from a model family that builds none of the agents read every session blind.
Where this comes from
Armature ran 15,051 judged sessions with Claude Code, Codex, Cursor, Grok Build CLI and Muse Code inside 92 realistic codebases, and published every run. The numbers on this page come from that work.