# Do coding agents recommend Cloudflare R2? > Cloudflare R2 was chosen in 13% of 90 judged file storage sessions, ranking fourth. Measured with Claude Code, Codex and Cursor. Source: https://armature.tech/library/do-coding-agents-recommend-cloudflare-r2 Published: 2026-09-03 Publisher: Armature, Inc. (https://armature.tech) --- > Cloudflare R2 was chosen in 13% of 90 judged file storage sessions, ranking fourth. It was also raised as a candidate in 39 further sessions without being chosen. This page reports what happened when Claude Code, Codex and Cursor had to solve a problem in file storage inside a realistic codebase. Not what a chat assistant says about Cloudflare R2. What an agent actually installed. ## The numbers | | | | --- | --- | | Category | File storage | | Sessions in the category | 90 | | Sessions where Cloudflare R2 was chosen | 12 | | Install share | 13% | | Rank in category | 4 of 5 | | Codebases it won in | 4 | | Raised as a candidate, not chosen | 39 | | Chosen when considered | 24% | | Site | cloudflare.com | ## By agent With 12 wins spread across three 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 Cloudflare R2 | Share | | --- | --- | --- | --- | | Claude Code | 30 | 8 | 27% | | Codex | 30 | 1 | 3% | | Cursor | 30 | 3 | 10% | ## By who was asking Cloudflare R2 does much better with one kind of buyer than another. It won 28% of sessions asked as junior developer and 0% of those asked as enterprise team. | Who is asking | Sessions | Chose Cloudflare R2 | Share | | --- | --- | --- | --- | | Junior developer | 18 | 5 | 28% | | Senior engineer | 27 | 7 | 26% | | Enterprise team | 36 | 0 | 0% | ## What Cloudflare R2 was up against The full ranking in file storage, from the same sessions: | # | Product | Runs won | Share | | --- | --- | --- | --- | | 1 | Amazon S3 | 41 | 46% | | 2 | Azure Blob Storage | 18 | 20% | | 3 | Google Cloud Storage | 18 | 20% | | 4 | Cloudflare R2 **(this page)** | 12 | 13% | | 5 | Tigris | 1 | 1% | ## What this means This is an integration problem, not a presence problem. Agents raised Cloudflare R2 in 39 sessions and chose it in 12, 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 90 sessions in file storage are part of a published set of 5,292, run with real coding agents inside realistic codebases and judged blind. The full method is on one page: [how we measured this](/library/how-we-measured-this). Every file storage run can be replayed on [the board](/leaderboards/storage). If you work on Cloudflare R2: the judge recorded a reason for every session where it was raised and passed over. Those reasons are in the transcripts. ## Common questions ### Do coding agents recommend Cloudflare R2? Yes. Cloudflare R2 was chosen in 12 of the 90 judged sessions in file storage, a 13% install share, ranking fourth in its category. ### Does Claude Code recommend Cloudflare R2? In 8 of the 30 sessions in file storage run with Claude Code, which is 27%. ### Do different coding agents treat Cloudflare R2 differently? Yes, and by a wide margin. Claude Code chose it in 27% of its runs and Codex in 3%. ### How was this measured? Real coding agents at pinned versions were run in sandboxes inside 51 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 Cloudflare R2 considered but not chosen? It was raised as a candidate in 39 sessions without being chosen, and chosen in 12. That is a 24% conversion from considered to chosen. ## Read next - [How to get picked for file storage by coding agents](https://armature.tech/library/storage-coding-agents-playbook) (Markdown: https://armature.tech/library/storage-coding-agents-playbook.md) - [Do coding agents recommend Amazon S3?](https://armature.tech/library/do-coding-agents-recommend-amazon-s3) (Markdown: https://armature.tech/library/do-coding-agents-recommend-amazon-s3.md) - [Do coding agents recommend Azure Blob Storage?](https://armature.tech/library/do-coding-agents-recommend-azure-blob-storage) (Markdown: https://armature.tech/library/do-coding-agents-recommend-azure-blob-storage.md) - [Do Claude Code, Codex and Cursor pick the same tools?](https://armature.tech/library/do-claude-code-and-codex-agree) (Markdown: https://armature.tech/library/do-claude-code-and-codex-agree.md) --- Armature helps software products get discovered and used by coding agents. Service: https://armature.tech/discoverability · Results: https://armature.tech/leaderboards/sectors · Contact: contact@armature.tech