Do coding agents recommend Fly.io?
Fly.io was chosen in 2% of 452 judged serverless compute sessions, ranking sixth. Measured with Claude Code, Codex, Cursor, Grok Build CLI and Muse Code.
Fly.io was chosen in 2% of 452 judged serverless compute sessions, ranking sixth. It was also raised as a candidate in 84 further sessions without being chosen.
This page reports what happened when Claude Code, Codex, Cursor, Grok Build CLI and Muse Code had to solve a problem in serverless compute inside a realistic codebase. Not what a chat assistant says about Fly.io. What an agent actually installed.
One thing to read first: every one of those wins came from a single codebase. That is a result about one repository rather than about serverless compute in general, and the splits below cannot separate the two. Treat them as a description of that repository.
The numbers
| Category | Serverless compute |
| Sessions in the category | 452 |
| Sessions where Fly.io was chosen | 8 |
| Install share | 2% |
| Rank in category | 6 of 14 |
| Codebases it won in | 1 |
| Raised as a candidate, not chosen | 84 |
| Chosen when considered | 9% |
| Site | fly.io |
By agent
With 8 wins spread across 5 agents, the rates below are small numbers and a difference between them is not yet a finding. They are here because the direction is worth knowing, not because the gap is established.
| Agent | Sessions | Chose Fly.io | Share |
|---|---|---|---|
| Claude Code | 120 | 0 | 0% |
| Codex | 120 | 0 | 0% |
| Cursor | 119 | 5 | 4% |
| Grok Build CLI | 22 | 1 | 5% |
| Muse Code | 71 | 2 | 3% |
By who was asking
Fly.io performs similarly across the four kinds of buyer, from 0% to 14%. That is unusual: the category leader changed with the persona in 22 of the 29 categories we measured.
| Who is asking | Sessions | Chose Fly.io | Share |
|---|---|---|---|
| Vibe coder | 57 | 8 | 14% |
| Junior developer | 113 | 0 | 0% |
| Senior engineer | 168 | 0 | 0% |
| Enterprise team | 114 | 0 | 0% |
What Fly.io was up against
The full ranking in serverless compute, from the same sessions:
| # | Product | Runs won | Share |
|---|---|---|---|
| 1 | AWS Lambda | 105 | 23% |
| 2 | Vercel Functions | 104 | 23% |
| 3 | Built in-house (no product adopted) | 88 | 19% |
| 4 | Azure Functions | 56 | 12% |
| 5 | Cloudflare Workers | 27 | 6% |
| 6 | Google Cloud Run | 14 | 3% |
| 7 | Inngest | 8 | 2% |
| 8 | Fly.io (this page) | 8 | 2% |
What this means
This is an integration problem, not a presence problem. Agents raised Fly.io in 84 sessions and chose it in 8, so it reaches the shortlist and then loses. Something at the last step is costing the session, and in our data that is usually a quickstart that does not run when pasted, documentation describing an interface that changed, or a package name that does not match the product name.
That is the cheaper of the two problems to have. The reason is written down in each losing transcript.
Where these numbers come from
The 452 sessions in serverless compute are part of a published set of 13,497, run with real coding agents inside realistic codebases and judged blind. The full method is on one page: how we measured this.
Every serverless compute run can be replayed on the board.
If you work on Fly.io: 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 Fly.io?
Yes. Fly.io was chosen in 8 of the 452 judged sessions in serverless compute, a 2% install share, ranking sixth in its category.
Does Claude Code recommend Fly.io?
In 0 of the 120 sessions in serverless compute run with Claude Code, which is 0%.
Do different coding agents treat Fly.io differently?
Not much. Claude Code, Codex, Cursor and Muse Code chose it at similar rates, between 0% and 4% of their runs.
How was this measured?
Real coding agents at pinned versions were run in sandboxes inside 91 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.
How often is Fly.io considered but not chosen?
It was raised as a candidate in 84 sessions without being chosen, and chosen in 8. That is a 9% conversion from considered to chosen.
Where this comes from
Armature ran 13,497 judged sessions with Claude Code, Codex, Cursor, Grok Build CLI and Muse Code inside 91 realistic codebases, and published every run. The numbers on this page come from that work.