Search & Discovery
Editorial Research

By · Published · Updated

The French Developer Who Built a Pocket Alternative Before Pocket Disappeared

How Nicolas Lœuillet's worry about Google Reader became wallabag and why its moment arrived just as Mozilla pulled the plug on its own read-later service.

Key Takeaways · Quick Answers
What is wallabag and when did it start?
Wallabag is a self-hostable, open-source read-it-later application. It began in 2013 as a project called "poche" (French for pocket), created by Nicolas Lœuillet after Google announced it was shutting down Google Reader. The project was renamed to wallabag in 2014 due to trademark considerations with Pocket. It has been MIT-licensed since its inception.
How is wallabag different from Pocket or Instapaper?
Wallabag is open-source and self-hostable, meaning you can run it on your own server and control your own data. Pocket and Instapaper are proprietary services owned by companies; Mozilla acquired Pocket in 2017 and shut it down in July 2025. Wallabag has no advertising, no tracking, and no data resale its hosted service (wallabag.it) is funded by subscriptions. The MIT license means the code is always available and can never be taken away from the community.
Can I import my saved articles from Pocket to wallabag?
Yes. Wallabag supports importing from Pocket CSV exports, as well as from Omnivore, Instapaper, and Pinboard. Version 2.6.14 (released October 2025) included Pocket CSV import functionality, and earlier 2025 releases also addressed Pocket migration needs. The wallabag migration documentation provides step-by-step instructions for each import path.
What does it cost to use wallabag?
Wallabag itself is free to download and self-host on your own server. The hosted service, wallabag.it, starts from €11 per year with a 14-day free trial that requires no credit card. There are no ads, no data resale, and no feature paywalls within the subscription tiers. You can export all your data anytime.
Who maintains wallabag?
Wallabag is maintained by Nicolas Lœuillet, Thomas Citharel, Jérémy Benoist, Kevin Decherf, Yassine Guedidi, and many community contributors. The project has over 12,000 commits on GitHub and nearly 13,000 stars. The community maintains translations, documentation, and additional client applications. The team holds regular informal meetups (apéros wallabag) in Paris and participates in community UX sessions like those organized by La Clinique du Libre.

In August 2013, Google announced it was killing Google Reader.

The RSS feed aggregator had become a quiet essential for people who followed the web through its own output more than its algorithmic feeds. When Google pulled the plug, the disappearance felt sudden and total years of carefully curated subscriptions, gone in a product sunset announcement. For most users, it was an inconvenience. For Nicolas Lœuillet, a French developer, it was something closer to an alarm.

If Google could do that to Reader, what was stopping it from doing the same to Pocket?

"This alarmed Nicolas Lœuillet, who worried that the same thing might happen to Pocket," reports Joe Brockmeier at LWN.net. "To ensure he had a home for his saved articles and data he began work on a project for self-hosting saved web content called 'poche', which is French for 'pocket'."

That worry, crystallized into code, is the origin story of wallabag a read-it-later application that would quietly accumulate features, contributors, and philosophical commitments over the following decade, until its moment arrived in the summer of 2025.

A Rename Before the Name Stuck

The choice of "poche" was practical and personal: it described exactly what the application did, and it was a word Lœuillet already lived with. But by 2014, the project needed a new identity. Pocket, the popular read-later service that had rebranded from Read It Later back in 2012, had its own trademarks to protect.

"He renamed it to wallabag in 2014 following some trademark unpleasantness," Brockmeier notes, using a word that carries the same internal logic as its predecessor a container for things worth keeping while removing the collision.

The rename marked an early sign of the project's character: pragmatic about the constraints of the broader software ecosystem, unwilling to fight unnecessarily, and focused on the work itself. Wallabag continued adding features while respecting the boundaries of the landscape it operated in.

Building the Stack: Symfony, MIT, and the Long Game

The technical architecture of wallabag reflects choices that have aged remarkably well. The core application is written using the Symfony PHP framework, a decision that provided structure, maintainability, and a familiar development environment for contributors who came to the project over the years. The project operates under an MIT license, meaning anyone can run it, modify it, fork it, or deploy it commercially without the constraints of copyleft licensing.

"The project consists of a web application written using the Symfony PHP framework," LWN.net confirms, "as well as a number of client applications and browser extensions to save data to wallabag or fetch articles for reading."

By 2025, wallabag had accumulated a substantial ecosystem of clients and extensions. Firefox and Chrome users could save pages directly through official browser extensions. The project offered official Android and iOS clients (with the Android app licensed under GPLv3 more than MIT), and the community had produced a range of additional options: a GNOME application called Read It Later, a command-line client, an emacs client, and several e-reader integrations for reading saved content on dedicated devices.

For developers wanting to build their own integrations, the wallabag API has "API wrappers in Go, Java, JavaScript, and Rust," according to the LWN.net coverage. The API is documented well enough that building a custom client is a realistic option beyond a theoretical one.

The Hosted Service: wallabag.it and European Infrastructure

Self-hosting has always been the foundational promise of wallabag run it on your own server, control your own data. But not everyone wants to manage a server, and the project has always recognized that reality. In 2016, Lœuillet launched wallabag.it as a hosted service, offering managed accounts with professional support, automatic upgrades, and daily backups.

"wallabag has existed since 2013 and wallabag.it since 2016," the wallabag.it site explains. "The service is funded by subscriptions, not by ads, tracking or data resale."

The hosted service operates on European infrastructure, which matters to users in jurisdictions with strong data protection frameworks. Subscription funding removes the incentive for behavioral surveillance there's no advertising engine optimizing for engagement, no data being packaged and sold, no dark patterns designed to maximize time-on-platform. What you save stays yours, and wallabag.it's business model doesn't require knowing anything about what you read.

The pricing structure reflects this straightforward approach: plans start from €11 per year, with a 14-day free trial that requires no credit card. "No limitation and no credit card required," the wallabag.it page notes. The trial is genuinely trial-like an opportunity to evaluate the service beyond a bait-and-switch designed to capture payment details.

For users who later decide to leave, the exit path is equally clear. "Export your data anytime," the wallabag.it page promises. Data portability is built into the service more than treated as an afterthought or a premium feature.

The Migration Moment: When Pocket Shut Down

Mozilla acquired Pocket in 2017 and maintained it as a free, ad-supported service for years. Then, in May 2025, Mozilla announced it was shutting the service down. The closure would come in July 2025 roughly two months of notice for users who had accumulated years of saved reading.

The announcement triggered what must have felt like a familiar anxiety for anyone who remembered Google Reader: where do my saved articles go now? The CSV export was technically functional but philosophically hollow a spreadsheet of links, stripped of the full-text archive and organization that made a read-later service useful in the first place.

"When Pocket shut down with little notice, thousands of carefully curated articles vanished into a CSV export," wrote developer Pascal Cescato on DEV Community. "I needed a replacement fast, but this forced migration became an opportunity to build something better."

Cescato's account of the migration is notable for its candor. He tested 15 applications with ratings above 4.3 stars before finding wallabag. "Outdated interfaces, missing features, and clunky implementations were the norm," he found. Raindrop.io looked promising but had API limitations that ruled out the automation he wanted. Fabric offered AI-generated summaries at €4.67 per month a single feature in a subscription model that reminded him how quickly costs accumulate.

"Wallabag: The Open Source Solution," Cescato wrote. "Beneath that simple exterior lies surprising depth: Complete content preservation: Full articles, images, offline access. Advanced organization: Tags, favorites, annotations, highlighting. Cross-platform ecosystem: Native apps for iOS, Android, browser extensions, desktop applications. Robust API with webhooks: Built for automation, not as an afterthought. Self-hosting control: Total ownership of your data and infrastructure. Seamless Pocket import: Near-transparent migration preserving metadata."

He deployed wallabag using Docker alongside n8n for automation, Traefik as a reverse proxy, and PostgreSQL for the database. The entire stack ran on a €5.99 per month VPS. "The setup provides HTTPS with automatic Let's Encrypt certificates, persistent data storage, and complete isolation through containerization," Cescato explained.

What Pocket users found, upon arriving at wallabag, was a service built on the same underlying philosophy but with a different structural commitment. Wallabag would never be sold to an advertising company. It would never need to pivot to make its revenue numbers. It would never sunset because it didn't fit a corporate parent's strategy. It was, by design, independent of the pressures that killed Google Reader and would eventually kill Pocket.

The Community Behind the Code

Wallabag is not a solo project. The main wallabag site lists Nicolas Lœuillet, Thomas Citharel, Jérémy Benoist, Kevin Decherf, Yassine Guedidi and many contributors as the people behind the project. The credits page on GitHub acknowledges the long tail of community effort that any mature open-source project accumulates.

"An open source project built in the open, translated by contributors, and documented by its users," the wallabag.org site describes it. The translation work is significant: wallabag is available in numerous languages, reflecting contributions from users who wanted the interface in their own language and were willing to do the work to make that happen.

The documentation, too, has been community-maintained. As of June 2026, wallabag's news feed noted that the documentation had moved into the main wallabag code repository, simplifying contributions and tightening the connection between code changes and the docs that explain them.

The project holds regular informal meetups in Paris apéros wallabag, as the French phrasing puts it. These are social gatherings, not formal conferences, where the team meets users and talks about what they're working on. The wallabag news feed includes announcements for events like a July 2026 gathering on Wednesday, July 29, described as "Rendez-vous à Paris pour rencontrer l'équipe wallabag autour d'un verre" a chance to meet the wallabag team over a drink.

In June 2026, the wallabag team participated in a UX session organized by La Clinique du Libre, a community effort to improve open-source projects through user experience review. The wallabag news feed described the session as "une heure pour améliorer le projet" an hour spent improving the project focused on practical usability improvements more than cosmetic changes.

These details matter because they describe a project culture beyond just a code repository. Wallabag is maintained by people who engage with its users in person, who accept UX help from the community, and who communicate in French in their home context while maintaining English documentation for international users.

What This Means for Lnk2It Readers

Link curation is, at its core, an act of future-proofing: saving something today that you want to find tomorrow. The services we use to do that curation carry risks that are easy to overlook until the service disappears.

Wallabag's architecture addresses that risk at its root. When you self-host wallabag, your reading archive lives on infrastructure you control. When you use wallabag.it, your subscription funds a service with no advertising incentive to surveil your reading habits. The MIT license means the code will always be available for others to run, fork, or fork into a new project if the current maintainers ever move on.

For readers researching link curation and resource discovery, wallabag represents a working example of what an open-source read-it-later service looks like when designed around data ownership more than data extraction. The migration story from Pocket documented in detail by developers like Pascal Cescato shows what the transition actually looks like in practice, including the Docker deployment details and the automation workflows that become possible when your data lives somewhere you fully control.

The 2025 Pocket shutdown was a loss for users who had trusted Mozilla to maintain a service it acquired. It was also a validation for wallabag, which had been quietly building an alternative for over a decade, betting that people would eventually want somewhere to save things that wouldn't vanish when a company's priorities shifted.

Version 2.6.14 and the Road Behind

As of October 2025, wallabag had reached version 2.6.14, a release that included Pocket CSV import functionality. The wallabag news feed records the release as part of a series of updates throughout 2025 that addressed compatibility with Pocket's export format infrastructure work done in response to the migration demand created by Pocket's shutdown.

Earlier in 2025, wallabab released version 2.6.13 with the same Pocket CSV import capability, and version 2.6.12 in April brought additional improvements. The release cadence reflects a project that had matured enough to respond to a surge in new users without breaking stride. No frantic pivots, no desperate fundraisers, no compromised privacy policies. Just code, updated, and released.

Version 2.6.10 appeared in November 2024, with version 2.6.9 following shortly after. The wallabag 2.7.0 branch was discussed in August 2023, suggesting that the team has been planning architectural improvements while maintaining the current stable release line.

On GitHub, the project shows over 12,000 commits and nearly 13,000 stars a measure of community interest that translates into real-world usage. The 895 forks represent people who have taken the code and adapted it for their own purposes, whether commercial deployments, customized instances, or experiments that may eventually feed improvements back into the main project.

The Honest Self-Hosted Migration Story

Self-hosting wallabag is not without friction. Pascal Cescato's DEV Community piece is valuable precisely because it doesn't pretend otherwise. Setting up Docker, configuring Traefik, managing PostgreSQL, and integrating n8n for automation requires technical comfort that not every reader-later user possesses.

But the alternative accepting that your reading archive lives entirely on someone else's servers, subject to their business decisions carries its own costs. Those costs are invisible until they aren't, until the acquisition, the pivot, or the shutdown notice arrives in an email.

Cescato's stack runs on a VPS that also hosts Nextcloud and other services. For him, wallabag fits into a broader self-hosted infrastructure philosophy a deliberate choice to reduce dependence on services that could disappear. "With Wallabag and n8n working together, I've built a system that surpasses what Pocket offered," he wrote, "for a fraction of the price and with complete control."

The "fraction of the price" is worth noting: a €5.99 VPS alongside subscription costs that might accumulate across multiple services with narrower feature sets. But the real value is control, not savings. The ability to export your data anytime, to run your own instance, to modify the code if you need something the official releases don't offer these are structural properties of open-source software that no amount of feature development can replicate for proprietary alternatives.

Where Wallabag Fits in the Read-Later Landscape

Wallabag occupies a specific position in the read-it-later market: open-source, self-hostable, MIT-licensed, with a hosted option for users who don't want to manage their own servers. It is not the only player in this space Instapaper, Pocket (until 2025), and Pinboard each serve overlapping needs but it is the only major read-later service with both a fully open codebase and a no-advertising business model.

"A read-it-later app designed for people who want ownership, not another closed inbox," wallabag.org describes itself. The phrasing is careful: not a dismissal of other services, but a statement of what wallabag is for.

The wallabag ecosystem supports imports from Pocket, Omnivore, Instapaper, and Pinboard practical migration paths that acknowledge users may arrive from elsewhere more than expecting them to start from scratch. The migration documentation walks users through each import process, recognizing that losing access to years of saved reading is not a hypothetical problem but a concrete one that actual users face.

For readers of Lnk2It who research tools, frameworks, and approaches to link curation, wallabag offers a case study in what open-source development looks like over a decade: the early decision to self-host, the community contributions that accumulated over years, the hosted service that provides income without compromising the core philosophy, and the moment when external events validated the entire approach.

Why This Matters

The story of wallabag is ultimately a story about trust and infrastructure. We trust services with our reading habits, our saved articles, our annotated notes and highlights. Those services are, for most users, invisible infrastructure unnoticed until something changes.

Nicolas Lœuillet noticed. In 2013, watching Google Reader disappear, he noticed that the services we depend on are not permanent. He built something that would be permanent in the way that matters: open-source, self-hostable, exportable, and independent of any single company's continued interest.

Twelve years later, when Pocket shut down, that work was ready. Not because wallabag had grown into a massive consumer brand it hadn't but because it had maintained its core commitments through a decade of releases, contributions, and community engagement. The project was there when people needed it, and it was there on terms that respected their data.

That is, perhaps, the most useful thing wallabag demonstrates: that building for permanence is possible, that open-source projects can sustain themselves over years, and that the services we depend on don't have to disappear when a company's strategy shifts. The code is on GitHub. The hosted option is available. The migration paths are documented. The project continues.

Where to Read Further

  • Joe Brockmeier's "Out of Pocket and into the wallabag" at LWN.net comprehensive technical overview and origin story from May 2025
  • Pascal Cescato's "From Pocket to Wallabag: A Self-Hosted Migration Story" on DEV Community detailed Docker deployment walkthrough and automation workflow
  • The official wallabag site features, migration guides, and current release information
  • Wallabag.it hosted service managed accounts with European infrastructure and 14-day free trial
  • Wallabag on GitHub source code, issue tracker, and contributor information
  • Wallabag news feed project updates, release notes, and community events

Sources reviewed

Atlas Research Network