What Pocket actually does
Pocket was a read-later and content curation app originally founded as Read It Later in 2007. Mozilla acquired it in 2017. On May 22, 2025, Mozilla announced Pocket's shutdown — immediately disabling new sign-ups and removing the app from stores. Monthly subscriptions were cancelled. On July 8, 2025, the service officially shut down, with users entering an export-only mode. All user data was permanently deleted by October 8, 2025. The Pocket API was disabled November 12, 2025.
Mozilla also shut down Fakespot (acquired 2023) in the same period: its Review Checker integration was killed June 10, 2025, and the apps/web went offline July 1, 2025. The newsletter 'Pocket Hits' was rebranded as 'Ten Tabs' and continues as a Mozilla curated email.
The shutdown prompted migration discourse dominated by Instapaper, Raindrop.io, and Readwise Reader. The read-later category contracted significantly in 2024–2025 — Omnivore also shut down on November 15, 2024 after an ElevenLabs acquihire. Karakeep (formerly Hoarder, 25.2k stars, AGPL-3.0) and Wallabag (12.7k stars, MIT) remain the primary self-hosted options, representing the only way to guarantee long-term data survival in this category.
Save-for-later with web content extraction
The core workflow: user clicks 'save to Pocket' from a browser extension or mobile share sheet, Pocket fetches the article and extracts clean readable text using Mozilla's Readability.js library.
Offline reading across devices
Saved articles sync to all devices with offline reading support. Content is stored locally on the device for reading without an internet connection.
Tagging and organization system
Users tag saved items for later organization. Tags were the primary organizational layer — no folders or hierarchical structure.
Full-text search across saved articles
Search by keyword across all saved article content. Premium feature in the original product.
Text-to-speech for articles
Pocket's Pocket Premium included text-to-speech playback for listening to saved articles. Mobile-focused feature.
Curated discovery feed and recommendations
Pocket Hits newsletter and curated discovery surfaces trending articles to users. Mozilla rebranded this as 'Ten Tabs' and continued it after the app shutdown.
Pocketpricing & limits
Pocket shut down July 8, 2025. All user data permanently deleted October 8, 2025. API disabled November 12, 2025. Annual subscribers received prorated refunds.
Where Pocket falls short
Complete platform shutdown with permanent data deletion
Pocket's July 8, 2025 shutdown gave users less than 6 weeks from announcement to shutdown for data export. Users who missed the October 8, 2025 export deadline lost all their saved articles, tags, and reading history permanently. Tens of millions of saved articles were deleted with no appeal process. The Pocket API was disabled November 12, 2025, breaking all third-party integrations built on it.
Mozilla's acquire-then-shut-down pattern destroyed user trust
Mozilla acquired Pocket in 2017 and shut it down in 2025 after 8 years. The same year, Mozilla shut down Fakespot (acquired 2023) — a product it had owned for less than 2 years. Omnivore also shut down in November 2024 after an ElevenLabs acquihire. The read-later category's 2024–2025 collapse demonstrated that depending on a free product owned by a cash-strapped organization is a data-loss risk.
No option to export after October 8, 2025 deadline
The export window was precisely 3 months from shutdown announcement (May 22) to data deletion (October 8). Users who were unaware of the shutdown, had inactive accounts, or were traveling during this period lost their data entirely. There was no extension offered and no open-source release of the codebase as an alternative.
Pocket API disabled — all integrations permanently broken
Third-party apps, browser extensions, and IFTTT/Zapier integrations built on the Pocket API were broken on November 12, 2025. Developers who had built tools on the Pocket API (Reeder, Instapaper import, productivity tools) had to rebuild connections to new services. This collateral damage extended Pocket's shutdown impact to the broader productivity tool ecosystem.
Curated content and Pocket Hits newsletter discontinued alongside the app
Pocket's curated content recommendation engine — which surfaced trending articles by topic — was shut down with the app. The newsletter 'Pocket Hits' was rebranded as 'Ten Tabs' but the app-based personalized discovery was discontinued. Users who relied on Pocket for content discovery (not just read-later) lost both the queue and the discovery mechanism.
Key features to replicate
The core feature set any Pocket alternative needs — plus what you can improve on.
Save-for-later with web content extraction (Readability)
Mozilla's Readability.js is open-source (Apache 2.0) and powers Pocket's content extraction. Given that Mozilla shut down Pocket, Readability.js is the cleanest available content extraction library to use in a custom build. It strips navigation, ads, and boilerplate HTML to extract the article body. Pair with Mercury Parser (archived but usable) or Postlight Parser as alternatives.
Offline reading across devices with cross-device sync
Content sync requires: storing extracted article HTML in the database, progressive web app (PWA) service worker for offline caching, and push sync when connectivity returns. For iOS/Android native apps, use React Native with SQLite for offline storage and a sync queue for writes made offline. This is the most complex feature — plan 2–3 weeks for reliable offline sync.
Tagging and organization system
Tags are simple string labels on saved items — implement as a tags table with an items_tags join table. Build a tag autocomplete UI, bulk-tag operations, and filter by tag. Add folder support as a hierarchical improvement over Pocket's flat tag model. Include favorites/starring as a secondary organization layer.
Full-text search across saved articles
Index extracted article content in Meilisearch or PostgreSQL full-text search (pg_trgm with GIN index). Search by title, author, domain, and article body. Add filter by tag, date saved, read/unread status, and domain. Full-text indexing at 10K articles is fast with PostgreSQL; scale to Meilisearch at 100K+ articles.
Text-to-speech for articles
Implement TTS using either browser Web Speech API (free, low quality) or a third-party TTS API (ElevenLabs, Google Cloud TTS, OpenAI TTS at $0.015/1K chars). Cache generated audio files in R2 keyed by content hash. Background audio playback in PWA requires the Media Session API for lockscreen controls.
Browser extensions and mobile apps
A browser extension (Chrome/Firefox Web Extension API) with a single click-to-save action is the primary save mechanism. The extension calls your platform's REST API with the current page URL and authentication token. For mobile, the iOS Share Extension and Android Share Intent capture URLs from other apps. Both require separate build and app store review processes.
Reading mode and customization
A distraction-free reading view renders Readability-extracted article HTML with: font size control, line height, font family (serif/sans), background color (light/sepia/dark), and estimated reading time. Store user preferences in the database and apply them consistently across all saved articles.
Technical architecture
A Pocket replacement is a read-later platform with content extraction, cross-device sync, full-text search, and offline reading. The core technical challenges are reliable content extraction (some sites block scrapers), offline sync consistency across devices, and browser extension distribution. The complexity is 6/10 — well-understood architecture with good OSS reference implementations in Karakeep and Wallabag.
Frontend (Web)
Next.js App Router, Remix, React SPA
Recommended: Next.js App Router with PWA configuration — enables offline reading via service worker, installable on desktop, and progressive enhancement for mobile.
Mobile Apps
React Native, Flutter, Capacitor (PWA wrapper)
Recommended: React Native — shares code with web frontend, native offline storage via SQLite (react-native-sqlite-storage), and native Share Extension for save-from-other-apps.
Content Extraction
Mozilla Readability.js, Mercury Parser, Diffbot
Recommended: Mozilla Readability.js (Apache 2.0) — the same library Pocket used; handles most article pages. Add Playwright headless browser fallback for JavaScript-rendered pages.
Database
PostgreSQL (Supabase), PlanetScale, SQLite (personal)
Recommended: PostgreSQL via Supabase — full-text search via pg_trgm for small scale, JSONB for extracted article metadata, RLS for user isolation.
Search
Meilisearch, Typesense, PostgreSQL full-text
Recommended: Meilisearch self-hosted — fast full-text search with typo tolerance; upgrade from PostgreSQL full-text at 100K+ saved articles.
Auth
Supabase Auth, Auth.js v5, Better Auth
Recommended: Supabase Auth — handles email, Google, Apple sign-in with RLS for multi-user article isolation.
File Storage
Cloudflare R2, AWS S3, Supabase Storage
Recommended: Cloudflare R2 — store cached article content, generated TTS audio files, and article images for offline reading.
Complexity estimate
Complexity 6/10 — the content extraction and offline sync are the technically interesting problems; everything else is standard CRUD. Plan for 8–10 weeks with 2–3 developers for a full MVP with browser extension and mobile app.
Pocket vs building your own
Open-source Pocket alternatives
Existing projects you can self-host or use as a starting point. Each has trade-offs.
Karakeep
25.2kKarakeep (formerly Hoarder) is a TypeScript bookmark and read-later manager (AGPL-3.0) with AI-powered automatic tagging, full-text search, multi-list organization, and browser extension. It is the most actively developed OSS Pocket replacement, with 25.2k stars and Docker-based self-host at 4/10 complexity.
Wallabag
12.7kWallabag is a PHP-based self-hosted read-later application (MIT) that has been actively maintained since 2013. It offers article saving, tags, full-text search, annotations, and export to EPUB/PDF. It is the most mature OSS Pocket replacement.
Omnivore (archived)
16.1kOmnivore was a TypeScript read-later platform (AGPL-3.0) that shut down in November 2024 after ElevenLabs acquihired the team. The codebase is archived on GitHub and self-hostable. It featured highlights, annotations, Obsidian/Logseq integration, and newsletter inbox.
Build vs buy: the real math
8–10 weeks
Custom build time
$15K–$30K
One-time investment
Karakeep self-hosted at $5/mo saves $60/yr versus Pocket Premium — custom build recoups in 250–500 years against SaaS savings
Breakeven vs Pocket
Pocket was free, so there is no SaaS cost to recoup. The build case is entirely about data sovereignty and platform survival — which Pocket's shutdown proved is a real, not theoretical, risk. Karakeep self-hosted at $5/mo total cost is the correct answer for 95% of use cases. A $15K–$30K custom build is only justified for: teams building a read-later product as a business (charging users for a personal knowledge management subscription), embedding article saving in a larger productivity app, or enterprises with compliance requirements needing private self-hosted article storage. For individuals and small teams: Karakeep or Wallabag self-hosted.
DIY roadmap: build it yourself
This roadmap builds a Pocket replacement with article saving, Readability extraction, tagging, search, and a browser extension. It assumes 2 developers using Next.js and Supabase, building on or inspired by Karakeep's architecture.
Article saving and extraction
2–3 weeks- Set up Next.js App Router with Supabase auth (email + Google OAuth)
- Integrate Mozilla Readability.js for server-side article extraction
- Build article save API endpoint: accept URL, extract, store in PostgreSQL
- Add Playwright headless browser fallback for JavaScript-rendered pages
- Build browser extension (Chrome/Firefox) with one-click save to your API
Organization and search
2–3 weeks- Build tag management with autocomplete and bulk-tag operations
- Implement full-text search with Meilisearch (index title, author, body)
- Add filter by: tag, domain, date saved, read/unread, starred
- Build distraction-free reading view with font/size/theme preferences
- Implement Pocket CSV import to migrate existing user data
Offline reading and sync
2–3 weeks- Add PWA manifest and service worker for offline reading support
- Implement offline article caching: store extracted HTML in service worker cache
- Build sync queue for actions taken while offline (save, tag, mark read)
- Add estimated reading time calculation (words / 200 WPM)
- Build full data export (JSON + EPUB options) for user data portability
These estimates assume 2 experienced developers. Building browser extensions for Chrome and Firefox requires separate manifest configurations (Manifest V3) and app store review for distribution. React Native mobile apps add 4–6 weeks beyond this roadmap. Consider forking Karakeep instead of building from scratch — its TypeScript codebase is well-structured and AGPL-3.0.
Features you can't get from Pocket
This is where a custom build pulls ahead — features impossible or impractical on a shared platform.
AI-powered automatic tagging and content categorization
Karakeep already includes this — using an AI model to automatically tag saved articles by topic, domain, and content type without manual effort. A custom build can extend this with personalized topic clusters based on reading history, surfacing 'similar articles you've saved' and 'topics you haven't read recently.'
Newsletter inbox with automatic article extraction
Omnivore had a dedicated email address that converted newsletter emails to read-later articles. A custom platform can provide a personal @yourplatform.com email address; incoming emails are parsed and saved as articles in the user's queue. This was Omnivore's most popular feature — and is not available in Karakeep or Wallabag.
Highlight and annotation system with PKM export
Pocket had no highlights. A custom platform can add text selection highlights with colored labels and personal notes, exportable to Obsidian (Markdown), Notion (API), or Logseq. This transforms a simple read-later tool into a knowledge capture system — the gap that Readwise Reader fills at $120/yr.
Reading list sharing and curation for teams
Pocket was personal. A custom platform can add team reading lists — shared queues where team members save articles and see what colleagues have saved. Add weekly digest emails of team reading activity. This turns a personal productivity tool into a team knowledge-sharing platform.
Who should build a custom Pocket
Individual users who lost data in the Pocket shutdown
Users who lost saved articles in Pocket's October 2025 data deletion need a self-hosted replacement where data deletion is controlled by the user, not a third-party organization. Karakeep or Wallabag self-hosted at $5/mo permanently prevents a repeat of the Pocket scenario.
Privacy-conscious researchers and journalists
Journalists and researchers saving sensitive articles (legal documents, investigative sources, sensitive research) need a private self-hosted article store where no third party can access or delete saved content. A self-hosted build with encryption-at-rest addresses this need that Pocket never could.
Knowledge management tool builders
Building a read-later layer into a PKM tool (Obsidian plugin, Notion workspace, Logseq integration) requires a custom API-driven article saving service. Karakeep's API is a good reference architecture, but a custom build allows deeper integration with the host PKM system.
Skip the DIY — let RapidDev build it
Everything above is doable — but it takes months of full-time work. We build custom Pocket alternatives using AI-accelerated development, delivering in weeks what used to take quarters.
Discovery call (free)
30 minWe map your exact requirements: which Pocket features you need, what custom features to add, your users, integrations, and compliance needs. You get a detailed scope document and fixed-price quote within 48 hours.
AI-accelerated build
8–10 weeksOur engineers use Claude Code, Lovable, and custom AI tooling to build 3–5x faster than traditional development. You see progress in a staging environment every week — not a black box for months.
Launch + handoff
1 weekWe deploy to your infrastructure, transfer the GitHub repo, set up CI/CD, and walk your team through the codebase. You own 100% of the source code — no vendor lock-in, no recurring platform fees.
What you get
Timeline
8–10 weeks
Investment
$15K–$30K
vs Pocket
ROI in Karakeep self-hosted at $5/mo saves $60/yr versus Pocket Premium — custom build recoups in 250–500 years against SaaS savings
30-min call. Fixed-price quote within 48 hours. No commitment.
Frequently asked questions
How much does it cost to build a Pocket alternative?
A custom Pocket replacement costs $15K–$30K with 2–3 developers. Karakeep self-hosted costs under $5/mo on a VPS and is the right answer for individuals. Build custom when adding team features, newsletter inbox, PKM integrations, or embedding read-later in a larger product.
How long does it take to build a Pocket clone?
Plan for 8–10 weeks with 2 experienced developers. Article extraction and saving take 2–3 weeks. Organization and search add 2–3 weeks. Offline PWA sync adds 2–3 weeks. Browser extension adds 1–2 weeks. React Native mobile apps add 4–6 weeks on top of the web app.
Are there open-source Pocket alternatives?
Karakeep (25.2k GitHub stars, AGPL-3.0) is the best active OSS alternative — TypeScript, AI tagging, browser extension, mobile apps. Wallabag (12.7k stars, MIT) is more mature with EPUB/PDF export. Omnivore (16.1k stars, AGPL-3.0, archived) is self-hostable but has no active development or security updates.
Can RapidDev build a custom Pocket alternative?
Yes. RapidDev has built 600+ apps including content management tools, browser extensions, and offline-capable PWAs. We can build a custom read-later platform with article extraction, offline sync, and highlight/annotation features. Book a free consultation at rapidevelopers.com/contact.
Can I recover my Pocket data after the October 2025 deletion?
No — all Pocket user data was permanently deleted by October 8, 2025 with no recovery option. If you exported before the deadline, your export file contains your saved articles list. If you missed the deadline, the data is gone. Mozilla offered no extension and no open-source release of the data or codebase.
Why did Mozilla shut down Pocket?
Mozilla has not published a detailed explanation. The shutdown follows a pattern with Mozilla: acquiring products (Pocket in 2017, Fakespot in 2023) and shutting them down years later as the organization refocuses. Mozilla has faced financial pressures with declining Firefox market share and revenue heavily concentrated in Google's default search agreement. Pocket was free for most users — building a sustainable revenue model around it proved difficult.
What happened to Omnivore and the read-later category?
Omnivore shut down November 15, 2024 after ElevenLabs acquihired the founding team. Pocket shut down July 8, 2025. Two of the three major free read-later apps disappeared within 8 months. Instapaper and Readwise Reader are now the primary commercial options. The category's collapse confirmed what self-hosting advocates argued: free tools dependent on external funding are not reliable for long-term data storage.
We'll build your Pocket
- Delivered in 8–10 weeks
- You own 100% of the code
- No per-seat fees, ever
30-min call. No commitment.