mrkeyoor.com_
Tue 01 Sept 17:40 UTC
Self-Hostedevaluationupdated 24 Aug 2026

puter review

Puter is an open-source, self-hostable desktop that runs in a web browser, bringing files, apps, games, and everyday utilities into one interface. It also gives developers hosted building blocks such as AI access, object storage, a key-value database, and serverless workers, plus an app store for publishing their work.

+79stars / 7d
Verdict

Puter is unusually ambitious and, based on our clean install and build, much more tangible than its grand internet-computer label might suggest. Use the hosted demo first, then self-host it if the integrated desktop and developer platform are genuinely useful to you. For conservative production adoption, the missing test target and large dependency footprint deserve a deliberate deployment and upgrade plan.

We ran it

Install✓ · 60s2944 packages · 1152 MB
Build✓ · 27s
Testsn/ano test script
Known vulns00 critical · 0 high · 0 moderate · 0 low (npm audit)
Repo1946 files~416,621 lines of source · 29.7 MB · 9 CI workflows · Dockerfile

Answers from our run

Does puter build from source?

Dependencies installed in 60 seconds (2944 packages), and the build succeeded in 27 seconds. We cloned commit 1ceffbe into a clean Debian container with 3 CPUs and no project-specific setup.

Does puter have tests you can run?

Not through a standard command: the project exposes no test script or target that our harness could run.

Does puter have known vulnerabilities in its dependencies?

npm audit found none in the dependency tree at the time of our run.

Who should not use puter?

Anyone seeking a small, single-purpose file server, because Puter is a full web desktop

What are the alternatives to puter?

Nextcloud, ownCloud Infinite Scale, CasaOS. Puter is unusually ambitious and, based on our clean install and build, much more tangible than its grand internet-computer label might suggest.

Setup4/5Install and build passed, but the dependency footprint is large
Docs4/5Clear entry points, self-host guide, and many translations
Community5/543,210 stars, fresh release, fresh push, and only 16 open issues
Maturity4/5Active releases and CI, tempered by no runnable test target

Discussed on

  1. hnShow HN: 3 years and 1M users later, I just open-sourced my "Internet OS"1,274 points
  2. hnShow HN: Firefox in WebAssembly273 points
  3. hnShow HN: Git Is Coming to Puter, "The Internet OS"4 points
  4. hnPuter Terminal: open-source browser terminal and shell with filesystem and AI4 points
  5. hnPuter 26.053 points

Who it’s for

People who want a browser-based personal cloud they can host themselves
Developers building web apps that need storage, data, workers, or distribution
Teams willing to operate a large TypeScript and Node.js application
Curious users who want to try the hosted service before maintaining their own instance

Who it’s NOT for

Anyone seeking a small, single-purpose file server, because Puter is a full web desktop
Operators who require a supplied automated test target before deployment
Organizations that cannot meet AGPL-3.0 obligations
Very small servers, because our npm install alone occupied 1,152 MB

Setup reality

In our fresh Debian sandbox, npm install succeeded in 60 seconds, installing 2,944 packages and using 1,152 MB on disk; the build then succeeded in 27 seconds. The repository had no test script or target, so tests were skipped. That is a smooth first build for a project of this size, but it is heavier than the README's short npm install and npm start sequence suggests, and the absence of a runnable test target leaves operators without an obvious pre-deployment confidence check.

What Puter is trying to be

Puter calls itself an open-source internet computer. In practical terms, it is a desktop-like environment delivered through the browser, with files, apps, and games collected under one roof. The README points to ordinary tools such as a notepad, voice recorder, spreadsheet, and camera, which makes the idea easier to understand: instead of visiting unrelated web tools, a user gets one workspace for everyday computing.

The project is also a platform for developers. Its documented services include AI access, object storage, a key-value database, and serverless workers. Developers can publish apps to Puter's app store, where the project says they can reach and monetize users. That combination distinguishes it from a conventional self-hosted file manager. Puter wants to own both the user-facing desktop and the services underneath third-party apps.

This is a large promise, but it is backed by a substantial repository rather than a concept page. Our checkout contained 1,946 files and roughly 416,621 lines of source. It is a TypeScript monorepo with npm workspaces, nine CI workflow files, a Dockerfile, and a Compose file. The AGPL-3.0 license supports inspection and self-hosting, while also requiring organizations to understand the license obligations before modifying or providing the software over a network.

What happened when we ran it

We cloned commit 1ceffbe into a fresh, unprivileged Debian container with Node.js 22, three CPUs, 8 GB of RAM, and no secrets. The npm install completed successfully in 60 seconds. It brought in 2,944 packages and occupied 1,152 MB on disk. That is a substantial dependency tree, but the important result is that installation finished without manual repair or undocumented system packages.

The build also succeeded, taking 27 seconds. npm audit reported zero known vulnerabilities across every severity level at the time of our run: zero critical, high, moderate, and low findings. That result is reassuring for this exact dependency snapshot, although it is not a security review or a promise about future advisories.

There was no tests script or test target, so we skipped tests. We also found no tests directory. This is the largest gap in the initial developer experience. A successful build proves the source compiled in our environment, but it does not establish that file operations, authentication, app installation, storage, workers, or desktop interactions behave correctly. Operators evaluating Puter should define their own smoke checks before upgrades.

Where Puter is strongest

The first strength is accessibility. The project offers a hosted service at Puter.com and links prominently to a live demo, so prospective users can judge the interface before running infrastructure. For self-hosters, the README supplies a shell installer for Linux and macOS, a PowerShell command for Windows, a separate self-hosting guide, and the familiar local-development path of clone, npm install, and npm start. The expected local address is stated explicitly as puter.localhost on port 4100.

The second strength is scope that remains coherent. The end-user apps and developer services support the same central idea: a browser environment where software and user data live together. Object storage, a key-value database, serverless workers, and an app store are not random extras. They are the pieces needed to make third-party apps feel native to the environment.

Weaknesses and rough edges

Puter's breadth is also its main cost. A browser desktop, cloud storage, app runtime, developer services, and distribution channel create a much larger operational and security surface than a focused file server. Even before persistent user data or production configuration, our node_modules footprint exceeded one gigabyte. Small machines, tightly controlled networks, and teams seeking a minimal service should account for that weight.

The README's quick-start commands are clear, but they do not communicate the architectural and operational decisions a serious deployment must make. Prospective operators still need to study the dedicated self-hosting documentation, plan backups, decide how upgrades are tested, and determine how public access is secured. The absence of a runnable automated test target makes that upgrade discipline more important.

The app platform proposition also creates dependency on Puter's own conventions. Developers should confirm that its storage, database, worker, publishing, and monetization paths fit their product before committing deeply. The README establishes that these facilities exist, but it does not provide enough detail in the truncated overview to judge quotas, compatibility boundaries, or migration paths. Those questions need documentation review and a prototype, not assumptions.

Community and project health

The repository has 43,210 stars, which indicates exceptional visibility, though stars alone do not prove support quality. More persuasive are the current activity signals: release 26.08.2 was published on August 24, 2026, and the repository was pushed again later that same day. With only 16 open issues reported, Puter does not present the usual picture of a famous repository buried under an enormous public backlog.

That combination suggests an actively maintained project with frequent shipping and strong adoption. It does not tell us response times or whether issues are closed by fixes, triage, or policy, so buyers should still inspect discussions relevant to their deployment. The supplied community item only confirms GitHub trending attention and adds no substantive user testimony.

Where it fits in a real stack

Puter fits best as a user-facing platform at the center of a personal cloud, lab environment, or web-app ecosystem. It is not merely a component to tuck behind an existing product. A realistic deployment surrounds it with the usual operational controls: TLS termination, access policy, monitoring, backups for user data, pinned releases, and smoke tests covering the workflows your users depend on.

For individuals, the sensible path is to try Puter.com, then self-host only if control over the environment justifies maintenance. For developers, build a small app against the documented storage, database, and worker interfaces before choosing the platform. For organizations mainly seeking synced files and collaboration, Nextcloud or ownCloud may be the more direct answer. Puter earns consideration when the integrated browser desktop and application platform are the point, not when they are extra complexity around a simpler need.

Alternatives

ProjectWhat it isPick it when
Nextcloud gh↗A mature self-hosted collaboration and file platform with a large app ecosystem.Pick this instead when file sync, sharing, calendars, and office collaboration matter more than a browser-desktop experience.
ownCloud Infinite ScaleA self-hosted file sync and sharing platform built around a modular server architecture.Pick this instead when your main requirement is organizational file access and sharing, not an internet computer or developer app platform.
CasaOSA home-cloud dashboard focused on installing and managing self-hosted applications.Pick this instead when you want a friendly home-server control panel for separate apps rather than one integrated web operating environment.

What people are saying

  1. [github-trending] HeyPuter/puter

Sources

  1. Puter GitHub repository
  2. Puter homepage

More self-hosted reviews

v2 · OpenShell · wigolo · Mindwtr · club-3090 · reclip · the whole board →