Readeck

Every read-it-later service asks for the same trade: hand over a running log of everything you read, and the service will hold your backlog for you. Readeck refuses that trade. It is a free, open-source web application that saves the readable content of web pages, including articles, long reads, pictures and videos. It strips out the ads and clutter and keeps the result indefinitely on a server its owner controls. That server can run on Linux, macOS, Windows or FreeBSD, or in Docker and Podman containers, which covers almost any machine a hobbyist already has running in a closet.

On iOS, the face of Readeck is a native client developed by Ilyas Hallak. It connects to a self-hosted instance through Readeck’s API and adds what a phone or tablet reader needs: dark mode, font customization, reading progress, a multi-column split view on iPad, offline-friendly syncing, and a share extension that clips links from Safari and other apps. A TestFlight beta runs alongside the main release for anyone who wants changes early. On desktop, a browser extension for Chrome and Firefox handles capture. Android users have a community-built app, ReadeckApp by jensomato, written in Kotlin with Jetpack Compose and distributed through F-Droid and GitHub rather than Google Play.

The market position is clear. Readeck shares a lane with Pocket and Wallabag and sits next to bookmark managers like Linkding, but it targets the self-hosting crowd specifically. That audience is small, and the store rating reflects it: 4.4 out of 5 from 13 ratings. The discussion that does exist, on Codeberg, GitHub and F-Droid, is positive. Users credit the clean reader view, labels and annotations, and the privacy of keeping everything at home. The recurring caveat is age. Readeck is the younger project, with a smaller ecosystem and fewer integrations than mature alternatives, and the iOS client is newly released.

• Readable-content extraction: Saves the article itself (text, pictures, video) without the ads and page furniture, and keeps it permanently on the owner’s server.
• Labels, favorites and archive states: Flexible tagging plus status flags that separate the active queue from finished and keeper material.
• Full-text search: Queries the saved content itself, not just titles and URLs.
• Annotations and highlights: Marks passages inside saved articles, turning the archive into source material.
• Text-to-speech: Converts the reading queue into audio.
• EPUB export: Packages articles for offline reading on e-readers and book apps.
• iOS share extension: Clips links from Safari and other apps through the system share sheet.
• iPad multi-column split view: A tablet-specific layout instead of a stretched phone interface.
• Reading customization: Dark mode, font controls and reading-progress tracking.
• Offline-friendly sync: Keeps saved content readable when the server is out of reach.
• Advanced connection options: Custom HTTP headers and local network access for non-standard self-hosted setups.

The Self-Hosting Bargain

Every judgement about Readeck starts with the server. The privacy argument holds up: when the instance runs on owned hardware, no third party sees the reading list, no business model depends on profiling it, and no company decision can wipe out years of saved articles. Data ownership here comes from the architecture, not from a promise in a privacy policy. That distinction matters most to people who have learned to distrust hosted services with their reading habits.

The bargain cuts both ways. Somebody has to install Readeck, keep it updated, back it up, and make it reachable from a phone. Broad OS support and container images lower that bar for anyone comfortable with Docker, but they do not remove it. The iOS client can do nothing for someone without an instance. It is a window onto a server, and without one it has nothing to show. That makes Readeck a poor recommendation for the general public and a strong one for the self-hosting community, and we judge it on those terms.

The Capture Loop

A read-it-later tool depends on how little friction sits between finding an article and saving it. Readeck covers the two places that matter. On desktop, the Chrome and Firefox extension saves the page in place. On iOS, the share extension hooks into the system share sheet, so a link in Safari, a news app or a message thread goes into the queue without a copy-and-paste detour. That is the correct design. Capture happens inside other apps, and a reading tool that forces a context switch before saving will lose links to impatience.

What Readeck saves matters as much as how it saves. It extracts the readable content instead of storing a URL and hoping the page survives. That is the difference between a bookmark and an archive. Link rot cannot touch an article already sitting on the user’s own disk, and the “keep them forever” pitch is credible because the content lives locally rather than in someone else’s cache. The saved copy is the product; the URL is only the starting point.

Organization and Retrieval

Saving is easy. Finding something six months later is hard, and that is where many reading queues turn into a guilt pile. Readeck gives users three retrieval tools. Labels provide overlapping categories without forcing a rigid folder tree. Archived and favorite states separate what is pending from what is done and from what deserves to stay close. Full-text search catches whatever the labels miss, because it searches the saved text itself. Together they make the archive something to consult rather than a graveyard.

Annotations and highlights push Readeck toward a research tool. Marking passages turns saved articles into source material, which suits students, writers, and anyone who reads to build an argument. Nothing here is exotic, since Wallabag and others offer comparable tools. But the set is complete, and completeness counts in a category where many free options stop at “save and read.”

Reading Beyond the Screen

Two features extend the queue past the app. Text-to-speech makes the backlog something to listen to on a walk or during chores. EPUB export sends long reads to e-ink devices and book apps, which suit a 40-minute feature better than a phone does. Both features address the real reason reading queues stall: people save more than they have screen time to read. Moving that content into other contexts is the fix, and Readeck builds that fix in rather than leaving it to third-party workarounds.

The Client Question

The iOS client’s advertised feature set focuses on reading comfort and sync: dark mode, fonts, reading progress, split view, offline-friendly syncing and the share extension. Annotations, text-to-speech and EPUB export are Readeck platform features, and the client’s listed strengths do not say how many of them appear inside the app. Anyone whose workflow depends on, say, highlighting on an iPad should confirm that before committing.

The client is also new, and the TestFlight channel signals active development. That is good for momentum, and a mild warning for anyone who wants a settled, finished tool. The advanced connection options, custom HTTP headers and local network access, show who the client is built for. They serve people whose instances sit behind reverse proxies or authentication layers, or on a home network that never touches the public internet. Those options are niche, but in a self-hosted product the niche is the audience. A client that ignored them would fail the people most likely to install it.

Readeck’s visual approach is subtraction. The standardized reading view discards the publisher’s layout, typography and ad furniture and renders every article in one consistent format. The trade-off is real: distinctive editorial design disappears along with the clutter, so a magazine feature and a personal blog post end up looking alike. For a tool whose job is focused reading, we think that trade is correct. Consistency lowers cognitive load, and the reader’s own choices, through font customization and dark mode, replace the publisher’s.

On iPad, the multi-column split view makes proper use of the larger display. Tablet readers that stretch a phone layout waste the screen, and this client avoids that mistake by design. Reading progress tracking addresses a small but persistent annoyance of long reads: coming back to a piece and hunting for where you stopped.

Performance on a self-hosted service depends mostly on the owner’s hardware and network, not on the app. A text-first reading client is light on device resources by design, and a self-hosted backend’s demands fall mostly on the owner’s hardware and network. The offline-friendly syncing should keep already-synced articles readable during a flight, a dead zone, or a server reboot. A self-hosted backend needs that protection more than a commercial cloud does.

• iPhone: Touch-first reading with system-level capture through the share extension. The phone is the natural place for collecting links throughout the day.
• iPad: The multi-column split view gives the tablet its own layout and makes it the stronger device for longer sessions and library management.
• Server dependency: The iOS app does not run standalone. It syncs with a self-hosted instance via Readeck’s API, so the iPhone and iPad experience is only as reliable as the owner’s server and network setup.
• Network flexibility: Local network access suits instances that live only on a home LAN. Custom HTTP headers accommodate proxies and gateways that expect extra request data.
• Pricing: Readeck itself is free and open source. The real cost is the hardware and time needed to host it.
• Release channels: A TestFlight beta sits alongside the main release, giving early adopters a path to newer builds.
• Cross-platform households: Android is served by a separate community app (ReadeckApp by jensomato) via F-Droid and GitHub, not Google Play. Two different developers maintain the two native clients, so feature parity between iOS and Android is not guaranteed.

Readeck makes a clear choice and follows through on it. It gives up the convenience of a hosted service in exchange for permanent, private ownership of everything a reader saves, and it backs that choice with a complete feature set. Content extraction, labels, full-text search, highlights, text-to-speech and EPUB export cover capture, retrieval and consumption without obvious gaps. Ilyas Hallak’s iOS client turns that backend into a credible reader, with a share extension in the right place, an iPad layout that respects the screen, and connection options that show an understanding of real self-hosted setups.

The limits are just as clear, and for most people they decide the outcome. The client is young, its exact parity with the web app is undocumented, the ecosystem is smaller than Wallabag’s, and the whole proposition collapses for anyone without a server. The standardized reading view also erases publisher design, and Android support depends on a separate community app. For committed self-hosters, Readeck is a strong fit; for everyone else, the setup burden and the unanswered questions make it hard to recommend. We recommend it to the first group and suggest the second look elsewhere.