The 7 Best Instapaper Alternatives in 2027 (When Minimalist Isn't Enough)

The best Instapaper alternative in 2027 is Marqly if the reason you are looking is that you save more than you read and can no longer find any of it — because the one thing Instapaper has never built is retrieval, and retrieval is the only gap that gets worse the longer you use the product. But that is one answer to seven different questions. If you want deeper reading and an annotation pipeline, it is Readwise Reader. If you want to listen rather than read, it is Matter. If you want an unlimited library for nothing, it is Raindrop.io. If you want to own the software, it is Wallabag, Karakeep or Linkwarden, all free to self-host.

Instapaper has always been the minimalist’s read-it-later app: save an article, read it in a clean, fast, text-only view, done. That simplicity is genuinely lovely, and after watching this category lose its two most-recommended tools inside eight months, its eighteen-year survival record is worth more than it was when we last wrote this page. But the simplicity is also the ceiling. There is no semantic search at any price, no automatic tagging, no summaries, no public sharing, no Edge extension and no browser-bookmark import. Free accounts cannot full-text search their own saves at all, and free notes cap at five per month. If you have hit that ceiling, here is the 2027 field.

Two things changed enough since our 2026 edition that this is a different list rather than a re-dated one. Pocket shut down on July 8, 2025 and Omnivore shut down in November 2024, so the two tools a 2026 version of this page might have pointed you toward are gone, and durability is now a ranking criterion rather than a nice-to-have. And every price below has been re-verified against a dated fact sheet — including Instapaper’s own, which our 2026 table got wrong. If you are weighing cost: Instapaper’s free tier covers unlimited saving, and Premium is $5.99/month or $59.99/year, per instapaper.com. Our full Instapaper review for 2027 has the plan-by-plan breakdown.

How this list was compiled. We compare reading experience, retrieval, AI features, offline support, archiving, portability, platform coverage and price, using dated competitor fact sheets and linked official sources. Every price and plan limit below carries the verification date of the sheet it came from. Marqly publishes this article and is one of the products discussed, ranked first because retrieval is the problem this query is about — where another tool wins a row against ours, the table says so. Our criteria are in how we test, and where we measured something rather than asserted it, the data lives in the import-fidelity study. We publish no numeric ratings for any product, including our own, and we invent no benchmark numbers.

What changed since our 2026 roundup

Nine changes, and the first two are corrections to our own numbers rather than to the market.

1. Our comparison table listed Instapaper at “$3/mo.” Premium is $5.99/month or $59.99/year, and has been since late 2023. Worse, the prose three paragraphs above that table quoted the correct price — so the 2026 edition contradicted itself on the page’s single most important number. That is the stale-price failure mode this category is full of, and it is why every figure below carries a verification date.
2. Readwise Reader was marked as having semantic search. It does not. Our fact sheet is explicit: “search is keyword-based rather than semantic.” Reader has Ghostreader AI for summarising documents, defining terms and answering questions about a text, plus full-text search and synced YouTube transcripts — which is a lot, and none of it is meaning-based retrieval. The 2026 table also priced it at “$12/mo”; the Full plan is $9.99/month billed annually, which is $119.88/year, or $12.99/month billed monthly.
3. Raindrop was priced at “$3/mo.” Pro is $28/year or $3.25/month — and the free plan, which is the reason to consider Raindrop at all, is unlimited bookmarks, collections and devices.
4. Marqly was priced at “~$6/mo.” Pro is $72/year (≈$6/month) or $9/month, the free plan stores up to 100 bookmarks with keyword search across the whole library, and every AI feature is Pro. There is also a standing $49 first-year offer — 31% off the $72 list price — and $48 in the first year for verified students.
5. Pocket and Omnivore are gone. Both are marked shutdown in our data layer. Pocket’s export window closed November 12, 2025; Omnivore’s closed November 30, 2024. Neither belongs on a 2027 alternatives list, and their absence reshapes the field: a lot of the libraries arriving at any of these tools are rescues rather than fresh starts.
6. Three tools we did not cover in 2026 are on the list now. Wallabag (open source, €11/year hosted, ePub export for e-reader owners), Karakeep (self-hosted AI with local Ollama support, semantic search added in 2026) and Linkwarden (automatic multi-format archiving plus Wayback Machine submission, with real team permissions). All three are free to self-host, which is a durability answer the 2026 edition did not offer.
7. Matter’s platform gap is now named. Matter has no Android app and no public API, and its App Store annual price is $79.99 against $60 on the web — an easy trap. The 2026 edition recommended it for listening without saying either.
8. Instapaper’s own feature set was under-described. It has a documented developer API, highlight sync out to Readwise, and export on the free plan. It does not have browser-bookmark HTML import, public sharing, or an Edge extension.
9. Durability is now a ranking axis. Two shutdowns in eight months means “will this exist in 2030” is a legitimate comparison column rather than a paranoid one. Section 11 below ranks the field on it.

When to leave Instapaper — and when to stay

Stay if all you ever want is to save an article and read it cleanly. Instapaper still does that better than anything on this list: eighteen years of typography and parser tuning, offline reading on iOS and Android, scheduled send-to-Kindle delivery, and a permanent archive on Premium. It is also cheaper than four of the seven alternatives below over five years. There is nothing wrong with simple, and there is nothing wrong with cheap.

Leave if you have noticed any of these:

• You save things and never find them again. This is the big one, and it compounds: the longer you use a retrieval-less tool, the larger the unsearchable pile gets.
• You want to search by what you remember, not by exact words. Instapaper’s search is keyword-literal, and it is Premium-only. Free accounts cannot full-text search their own saves at all.
• You want summaries to triage a backlog. Instapaper has no AI summarisation at any price.
• You want the filing done for you. Organisation is manual folders. No tags, no automatic collections, nothing resurfaced.
• You want your reading to become a searchable knowledge base. That requires a retrieval layer, and Instapaper does not have one.
• You need to publish a list to someone else — a class, a book club, a colleague, a newsletter audience. Instapaper has no public sharing of any kind.
• You listen more than you read and want good voices. Instapaper’s Premium text-to-speech is decent. Matter’s HD voices are the reason Matter exists.
• You want to own the software. After two shutdowns, this is a reasonable position, and Instapaper is closed-source and hosted.
• You live in Microsoft Edge. No Edge extension.
• You save things that are not articles. Videos, images, reference links and products all sit awkwardly in a queue built for prose.

If none of those landed, close this tab and keep Instapaper. It is a good tool and this page is not an argument that it is bad — it is an argument that “calm” and “findable” are different goals, and Instapaper has only ever pursued the first.

The 2027 comparison table

Marqly first because retrieval is what this query is about. ✅ / ❌ / ⚠️ / — as appropriate.

	Marqly	Instapaper	Readwise Reader	Matter	Raindrop.io	Wallabag	Karakeep	Linkwarden
Semantic search	✅ Pro	❌	❌ keyword	❌ keyword	❌ keyword	❌ keyword	✅ self-hosted	❌ keyword
AI summaries	✅ Pro	❌	✅ Ghostreader	❌	❌	❌	✅ BYO model	❌
AI auto-tagging	✅ Pro, every save	❌ folders only	❌ manual	❌ manual	⚠️ save-time suggestions on Pro	⚠️ rule-based	✅ BYO model	⚠️ optional, local models
AI Q&A over saves	✅ Ask, Pro	❌	✅ over the open text	❌	❌	❌	❌	❌
Free tier	⚠️ 100 bookmarks, keyword search	✅ unlimited saving	❌ none permanent	✅ unlimited library	✅ unlimited bookmarks	✅ self-host	⚠️ 10 bookmarks on Cloud	❌ none on Cloud
Paid price	$72/yr or $9/mo	$59.99/yr or $5.99/mo	$119.88/yr	$60/yr web · $79.99/yr iOS	$28/yr	€11/yr	$48/yr Cloud	$36/yr per user
Reading experience	⚠️ clean	✅ best in class	✅ power-reader	✅ best on iOS	⚠️	⚠️ functional	⚠️	⚠️ not the point
Text-to-speech	❌	✅ Premium	✅ iOS app	✅ HD, best in category	—	❌ lacks audio	—	❌
Offline reading	⚠️ Pro, web + iOS, per device	✅	✅	✅	❌	✅	✅ iOS, Android	✅ mobile caching
Automatic page archiving	⚠️ save as PDF on demand	⚠️ permanent archive, Premium	✅	—	⚠️ Pro	✅	✅ 3 formats	✅ 3 formats + Wayback
Public sharing	✅ boards, free	❌	❌	❌	✅	✅ article links	✅ lists with RSS	✅ collections
Self-host	❌	❌	❌	❌	❌	✅	✅	✅
Developer API	❌	✅	✅	❌	✅	✅	✅ REST + MCP	✅
Android	✅	✅	✅	❌	✅	✅	✅	✅
Safari extension	✅	✅	❌	❌	✅	❌	✅	❌
Edge extension	✅	❌	❌	❌	✅	❌	❌	✅
Fact sheet verified	Sep 26, 2026	Aug 2, 2026	Aug 2, 2026	Aug 2, 2026	Sep 12, 2026	Aug 2, 2026	Aug 2, 2026	Aug 2, 2026

A caveat on that table, because a grid like this invites over-reading. Every cell comes from a dated fact sheet rather than from a vendor’s marketing page or from memory, and the dates are in the bottom row — they are not all the same date, which means the sheets were not all refreshed at the same moment. Prices in particular move: Raindrop’s sheet is from September 2026 and Marqly’s from late September 2026, while the read-it-later sheets are from August 2026. Where a figure is not in a sheet we have written around it rather than filling the gap, and the two places that happens most visibly are Karakeep’s cloud annual price (not published, so we quote the monthly figure) and Instapaper’s free-plan advertising policy (not recorded, so we say nothing about it). For the same reason a — in any cell means our fact sheet records the capability neither way — not that we confirmed its absence. We would rather leave a cell empty than fill it from memory, which is exactly how the stale Instapaper price this page corrects got published in the first place. No cell in this table is a numeric rating, and we publish none for any product including our own.

1. Marqly — best for finding what you saved again

Marqly is the answer to the specific complaint that sends people away from Instapaper: you saved it, you half-remember it, and you cannot get it back. Semantic search on Pro finds saves by meaning across titles, page content, your highlights and YouTube transcripts, so “that piece about the economics of repair cafés” is a viable query rather than a lost cause. Automatic AI tagging runs on every save with no key to configure and no model to choose. AI summaries land on each item so you can triage a backlog without opening it. And Ask answers questions from your own library with numbered citations that open the source — the closest thing in this list to interrogating what you saved.

It is also the only tool here where the organisational work is done for you and the retrieval is meaning-based at the same time. Karakeep can do both, but only on your own server with your own model. Everything else on this list does one or neither.

The honest limits, because a comparison that only flatters its publisher is worthless: Marqly’s free plan is capped at 100 bookmarks and searches by keyword only — every AI feature is Pro. There is no public API, no self-hosting, and no text-to-speech. Offline reading exists but it is a Pro feature of the web and iOS apps only, and it is per-device with no server copy, so the offline library does not follow you between devices and the Android app and extensions do not have it. The reader is clean and less mature than Instapaper’s. And its browser-bookmark import is one-directional in a way that matters: it takes Chrome, Edge, Firefox and Safari HTML, Raindrop HTML, generic CSV, and the list.csv from inside a Pocket export archive — not Instapaper’s highlights, which no tool in this list imports intact.

Who it’s best for: readers whose Instapaper library has become a graveyard, and who would trade some typographic polish for being able to find things. Skip it if you want the best possible reading surface (Instapaper), the best audio (Matter), an annotation pipeline into Notion or Obsidian (Readwise Reader), an unlimited free library (Raindrop.io), or to own the code (Wallabag, Karakeep, Linkwarden).

• Pros: semantic search across titles, content, highlights and transcripts; automatic AI tagging on every save; AI summaries; Ask with cited answers; YouTube summaries, transcripts and chat; six-colour highlighting free on every plan; boards publish as public pages needing no signup to view, free; save as PDF screen-faithfully, processed locally in Chrome and Edge; ChatVault captures ChatGPT, Claude and Gemini conversations; widest platform coverage here — web, Chrome, Edge, Firefox, Safari, iOS and Android.
• Cons: free plan capped at 100 bookmarks; all AI is Pro; no public API; no self-hosting; no text-to-speech; offline is Pro, web and iOS only, per device; exact-URL deduplication only.
• Price: free plan up to 100 bookmarks, no card. Pro $72/year (≈$6/month) or $9/month. A standing $49 first-year offer — 31% off the $72 list price — and $48 in the first year for verified students, which is 33% off. Teams from three seats at $9/seat/month or $72/seat/year.

Get started free → · Marqly vs Instapaper · Instapaper migration guide

2. Readwise Reader — best for heavy readers who annotate

If Instapaper’s minimalism feels like a straightjacket rather than a relief, Readwise Reader is the other end of the spectrum. It folds articles, RSS feeds, email newsletters, PDFs, ePubs, Twitter threads and YouTube videos into one triage-oriented inbox, and its Ghostreader assistant uses AI to summarise documents, define terms and answer questions against the text. YouTube videos arrive with synced transcripts you can highlight like an article. It is keyboard-first, dense, and the fastest tool in the category once you have paid the learning cost.

The reason it is ranked second rather than first is the part that actually decides whether you should buy it: Reader is the annotation pipeline, not the retrieval layer. Everything you highlight flows into the core Readwise loop — daily spaced-repetition review, plus sync out to Notion, Obsidian, Evernote and Roam. If your reading is research and your output is notes in another tool, nothing here competes. If your problem is “I cannot find the article,” Reader’s search is keyword-based, its tagging is manual, and it will not fix that.

It is also the most expensive option on this list by a clear margin, and the pricing structure is unusual: Reader only ships with the Readwise Full plan, so you pay for Readwise highlight syncing whether you use it or not.

Who it’s best for: serious readers and note-takers consuming large volumes across articles, newsletters, PDFs and video, who want highlights flowing into Notion or Obsidian. Skip it if you are a casual reader (the UI is dense), you want a free tier (there is no permanent one), you want semantic retrieval, or you want to publish collections publicly.

• Pros: every content type in one inbox; Ghostreader AI summaries and Q&A over the open text; synced YouTube transcripts; best-in-class highlight workflow with daily review; exports to Notion, Obsidian and Roam; documented API; active, well-funded development.
• Cons: no permanent free tier; Reader requires the Full plan, so you buy highlight syncing whether or not you want it; dense UI with a real learning curve; no public sharing; tagging is manual; search is keyword-based rather than semantic; no Safari or Edge extension.
• Price: $9.99/month billed annually — $119.88/year — or $12.99/month billed monthly, after 30 days free. The cheaper Lite plan at $5.59/month billed annually covers classic Readwise highlight syncing without the Reader app. Readwise Reader vs Marqly has the row-by-row.

3. Matter — best for listening, if you are on Apple hardware

Matter is the design-forward read-later app for iPhone, iPad, Mac and the web, and it has the strongest text-to-speech in the category — human-like voices with word-by-word tracking, so you can switch between reading and listening mid-article without losing your place. That single capability is the reason to choose it, and it is genuinely better than Instapaper’s Premium narration.

It is more than an audio app. Parsing is excellent, newsletters arrive via Gmail integration or a dedicated Matter email address, and YouTube videos and podcasts are transcribed with time-synced text you can highlight. The free tier is unusually good: an unlimited library, share-sheet saving, a Chrome extension and unlimited tags. Premium adds HD voices, unlimited highlighting and notes, offline full-text search, and integrations with Notion, Obsidian and Readwise.

The disqualifiers are blunt and worth reading before you fall for the demo. There is no Android app. There is no public API and no public sharing of collections. And the App Store annual price is $79.99 against $60 on the web — a 33% markup for the convenience of Apple’s billing that a lot of subscribers never notice. Organisation is manual: no AI summaries, no chat, no automatic tagging.

Who it’s best for: Apple-ecosystem readers who listen as much as they read — commuters and audio-first users who want articles, newsletters and podcasts in one queue. Skip it if you use Android, you want an API, you want AI organisation, or you want to pay the lower annual price without noticing which store you subscribed through.

• Pros: the best listening experience in the category; beautiful, focused reading design on iPhone and iPad with Apple Pencil highlighting; newsletter ingestion via Gmail or a dedicated address; time-synced YouTube and podcast transcripts you can highlight; solid free tier with an unlimited library and unlimited tags; highlights sync to Notion, Obsidian and Readwise.
• Cons: no Android app; no public API; no public sharing; HD voices, unlimited highlighting and full-text search all sit behind Premium; no AI summaries, chat or automatic tagging; App Store annual billing costs $19.99 more than the web.
• Price: free tier with an unlimited library. Premium $8/month or $60/year on the web; $7.99/month or $79.99/year through the App Store. A higher Matter Pro tier at $12.99/month also appears in the App Store. Instapaper vs Matter.

4. Raindrop.io — best free library for visual collectors

Raindrop is the answer if the thing you resent about Instapaper is not the reading experience but the filing, and if you would rather not pay at all. Its free plan is the most generous hosted tier in the category: unlimited bookmarks, collections and devices, with up to three highlights per bookmark and 100 MB/month of uploads. No expiry, no card, no volume cliff.

It is a bookmark manager rather than a reader, which cuts both ways. It saves every media type — articles, images, videos, products — into structured collections with a visual polish Instapaper does not attempt, and it runs everywhere: web, Chrome, Edge, Firefox, Safari, iOS, Android, macOS, Windows and Linux. Ten platforms, the widest coverage on this list. Pro adds full-text search, permanent page copies, Stella AI save-time tag and collection suggestions, reminders, unlimited highlights in four colours with notes, annotations and daily backups.

What it does not have is an intelligence layer. Stella makes suggestions at save time; there is no automatic tagging applied to your library, no AI summaries, no semantic search and no chat. Retrieval still depends on your own tags and keywords — which is a better position than Instapaper’s, because at least the tags are yours to build, but not a fundamentally different one. And there is no offline reading mode.

Who it’s best for: collectors and visual organisers filing thousands of bookmarks into structured collections who want polished apps everywhere without paying. Skip it if you want a reading queue with typography (Instapaper or Matter), semantic retrieval (Marqly or Karakeep), offline reading, or automatic filing.

• Pros: the best free tier here — unlimited bookmarks, collections and devices with no expiry; ten platforms including every major browser and desktop OS; saves all media types; public and shared collections; permanent page copies on Pro; full-text search on Pro; documented API; imports from Pocket and browser bookmark HTML.
• Cons: no semantic search and no AI summaries or chat; Stella is save-time suggestion only, not automatic tagging; no offline reading mode; free plan limits highlights to three per bookmark and uploads to 100 MB/month; retrieval depends on your own tagging discipline.
• Price: free, unlimited. Pro $28/year or $3.25/month — App Store pricing is $27.99/year or $2.99/month — with yearly billing about 20% cheaper than monthly. See our full Raindrop.io review for 2027 and Marqly vs Raindrop.

5. Wallabag — best for ownership at the lowest possible price

Wallabag is the open-source read-it-later app that has been maintained continuously since 2013, started in France, and quietly outlived better-funded competitors. You can run it on your own server for free, or use the official hosted service at wallabag.it for €11 a year — among the cheapest paid plans anywhere in software, not just in this category. A €4 three-month option and a €30/year supporter plan round it out, and hosted accounts get 14 days free with no card.

The feature that wins it a permanent place on this list is ePub export, alongside PDF and other formats. If you read on a Kobo or another e-reader, Wallabag is the only tool here that sends your saves to the device natively, and it is why a lot of e-reader owners never leave. It also has rule-based automatic tagging — deterministic rather than AI, but genuinely hands-off — annotations on selected text, RSS feeds of your saves, public share links for individual articles, a full REST API, and importers covering Pocket, Instapaper, Readability and browser bookmark files. It is one of only two tools on this list that will import an Instapaper library directly.

The trade-off is stated plainly by its own fact sheet: the interface is functional rather than beautiful, parsing can fail or render poorly on complex, script-heavy sites, there are no AI features of any kind, and self-hosting requires real setup and ongoing maintenance across PHP, a database and updates.

Who it’s best for: privacy-focused readers and self-hosters who want a Pocket-style queue under their own control, plus e-reader owners who export to ePub. Skip it if polish affects whether you open the app, you want AI, or you have no appetite for running a PHP stack.

• Pros: open source and free to self-host; €11/year official hosting that funds development, with every feature on every plan; exports to PDF, ePub and other formats; rule-based automatic tagging; imports from Pocket, Instapaper and browser bookmark files; native iOS and Android apps with offline reading; public share links; full REST API; thirteen years of continuous maintenance.
• Cons: utilitarian design that feels dated next to commercial rivals; parsing can fail on complex, script-heavy sites; no AI of any kind — no summaries, no semantic search, no smart tagging; self-hosting means PHP, a database and ongoing updates.
• Price: free self-hosted. €11/year on wallabag.it, or €4 for three months, or €30/year for the supporter plan, with 14 days free and no card. Wallabag alternatives · Instapaper vs Wallabag.

6. Karakeep — best for self-hosters who want real AI

Karakeep — formerly Hoarder, renamed in early 2025 — is what happens when the self-hosted crowd builds an AI bookmark manager. Self-hosting is fully free and open source with no volume cap, and the AI is bring-your-own: point it at any OpenAI-compatible API key or at a local Ollama model, and it auto-tags and summarises everything you save. It added semantic and hybrid search modes during 2026 on top of a Meilisearch-backed full-text index, which makes it the only genuinely free path to meaning-based retrieval on this list.

The rest of the product is unusually complete for open source. Full-page archiving writes a page copy, a screenshot and a PDF for every link, with yt-dlp for video. Extensions ship for Chrome, Firefox and Safari — no Edge — and iOS and Android apps gained offline reading in 2026. Public lists have RSS feeds. And the automation surface is a REST API, webhooks, a CLI and an MCP server, so an AI agent can read and act on your library directly. Imports cover Pocket, browser bookmark HTML, Omnivore, Instapaper, Linkwarden and mymind — two of those six services no longer exist, which tells you who Karakeep is built for.

The costs are the self-hosted costs. You are the sysadmin, permanently: Docker, updates, backups you actually test, reverse-proxy security, and disk space that fills faster than you expect once every save writes three copies. The AI is free only on the Ollama path; a frontier API meters per thousand tokens. Karakeep is still pre-1.0. There is no AI Q&A over your saves and no YouTube-specific AI. And Karakeep Cloud, the hosted escape hatch, is in public beta with a $0 tier capped at 10 bookmarks and Pro at $4/month — no published annual price, and ten bookmarks cannot tell you what living with it is like.

Who it’s best for: self-hosters and privacy-focused tinkerers who want AI tagging, summaries and semantic search on their own hardware without a subscription. Skip it if you have no always-on machine, you live in Edge, you want to ask your library questions, or you want someone else to hold the pager.

• Pros: free, open source and self-hosted with all features and no caps; AI auto-tagging and summaries via any OpenAI-compatible API or fully local Ollama; semantic and hybrid search added in 2026; full-page archiving plus yt-dlp video; iOS and Android apps with offline reading plus Chrome, Firefox and Safari extensions; REST API, webhooks, CLI and an MCP server; imports from six sources including dead ones.
• Cons: self-hosting requires Docker and bring-your-own AI keys; Cloud is public beta with a 10-bookmark free tier and no published annual price; no AI Q&A over your saves and no YouTube AI; no Edge extension; no one-click save-all-tabs; pre-1.0 turbulence; archives fill disks.
• Price: free self-hosted. Karakeep Cloud: $0 tier capped at 10 bookmarks, Pro $4/month, public beta. See our full Karakeep review for 2027 and Marqly vs Karakeep.

7. Linkwarden — best for making sure the links still work

Linkwarden is the archivist’s answer, and it solves a problem none of the other six treat as central: the page behind your bookmark will change or disappear. Every saved link is automatically preserved as a screenshot, a PDF and a readable text copy, and Linkwarden can also submit it to the Wayback Machine. Instapaper’s permanent archive does something similar behind Premium; Linkwarden does it in more formats, self-hosted, for free.

It is also the most collaborative tool on this list by a distance. Collections nest, can be shared with teammates with per-member permissions, and can be published as public pages. For a research group, a content team, a classroom or a shared reading list, that is a capability Instapaper simply does not have. Self-hosting is free under AGPL-3.0 with all features included; Cloud is $3/month per user billed yearly — 25% off the $4/month monthly price — with up to 30,000 links per user, expandable via support, and no permanent free tier beyond 14 days.

The limits are structural rather than incidental. Search is keyword and full-text only — there is no semantic search. The AI tags and does nothing else: no summaries, no Q&A, nothing resurfaced. It is built for collecting and preserving rather than reading, with no reader queue, no text-to-speech and no reading progress. Extensions cover Chrome, Firefox and Edge — no Safari. Multi-format archiving means a headless browser doing captures and storage that grows fast, so a self-hosted instance wants a real server.

Who it’s best for: researchers, journalists, lawyers and citation-heavy writers who need pages as they were, plus teams sharing curated collections with permissions. Skip it if your problem is retrieval or triage, you read in your bookmark app, you live in Safari, or you want a free hosted account.

• Pros: automatic screenshot, PDF and readable copy for every link, plus Wayback Machine submission; AGPL-3.0 with genuine self-host parity and no held-back features; shared collections with per-member permissions and public pages; optional AI tagging including local models; full-text search across archived content; documented REST API; official iOS and Android apps with offline caching; $36/year per user on Cloud.
• Cons: no semantic search, no summaries, no Q&A; no reader queue, no text-to-speech, no reading progress; no free Cloud tier; no Safari extension; storage-hungry by design; mobile apps still catching up to the web.
• Price: free self-hosted. Cloud $3/month per user billed yearly (25% off) or $4/month billed monthly, 14 days free, up to 30,000 links per user. See our full Linkwarden review for 2027, Linkwarden vs Karakeep and Marqly vs Linkwarden.

Durability: two shutdowns reshaped this list

This is the section the 2026 edition could not have written, and it is the reason the ranking below differs from a pure feature comparison.

Mozilla shut Pocket down on July 8, 2025. Export access and the API were disabled on November 12, 2025, data deletion began then, and Premium subscribers received prorated refunds. Omnivore shut down in November 2024 after its team was acqui-hired by ElevenLabs, with an export deadline of November 30, 2024. Both are status: "shutdown" in our data layer. Pocket had been the default recommendation for a decade and was owned by a foundation. Omnivore was free, open source and widely beloved. Neither survived, and neither death was predictable from the product.

So “will this still exist” is now a legitimate column, and here is how the field scores on the four things we can actually verify — licence, revenue model, export path and import path:

Tool	Licence / code	Revenue model	Export you can rely on	Verdict
Instapaper	Closed	Premium since launch; independent since 2018	HTML or CSV on the free plan, developer API, Readwise highlight sync	Strongest survival record — 18 years, three owners, no data disaster
Wallabag	Open source	€11/yr hosted funds development	PDF, ePub and other formats; full REST API	Strongest ownership record — running continuously since 2013
Karakeep	Open source	Cloud, public beta	Karakeep JSON or standard Netscape HTML, lists included	Strong — a stalled project still leaves a working install
Linkwarden	AGPL-3.0, all features self-hosted	Cloud subscription	Standard bookmark HTML or JSON, downloadable archives, REST API	Strong — no open-core cripple, best portability here
Raindrop.io	Closed	Free tier + Pro	Bookmark HTML export; permanent copies on Pro	Good — mature, funded, but closed
Marqly	Closed	Free tier + Pro + Teams	CSV export button — URL, title, description, tags; complete account copy via support	Good — real self-serve export, but the CSV does not carry highlights, notes or boards
Readwise Reader	Closed	Readwise Full subscription	Highlights and notes sync out to Notion, Obsidian, Evernote, Roam; documented API	Good — the annotation export path is the best anywhere
Matter	Closed	Free tier + Premium	Highlights sync to Notion, Obsidian, Readwise; no public API	Weakest — no API and no public sharing means fewer ways out

Two honest notes on our own row. Marqly’s CSV export carries URL, title, description and tags — nothing else, so it does not carry highlights, notes or boards; for a complete account copy you email support@marqly.com. On the free plan that CSV covers your newest 100 items, and Pro exports the whole library with an option to export only what is new since the last export. And Marqly is closed-source with no self-hosted option, which means its durability rests on the company rather than on a licence. We are not going to claim the open-source tools’ advantage here.

The practical conclusion is not “pick open source.” It is test your exit before you need it. Export a small library on the day you adopt a tool, open the file, and check whether the thing you actually care about — links, highlights, notes, archives — is in it. Our import-fidelity study exists because we ran that test on ourselves: measured on September 26, 2026, 0 of 261 items parsed from a Pocket export’s HTML preview file and 43 of 43 parsed from a Chrome bookmark export. The file you feed an importer matters more than the importer’s reputation, and the same is true in reverse when you leave.

Which Instapaper alternative should you choose?

Match the pain, not the feature list:

• You save more than you read and cannot find anything again → Marqly. Semantic search, automatic tagging, summaries and Ask are all aimed at exactly that failure. Marqly vs Instapaper.
• You read a hundred things a week and your output is notes in Notion or Obsidian → Readwise Reader. The annotation pipeline and spaced repetition are unmatched; accept $119.88/year and the learning curve.
• You would rather listen than read → Matter if you are on Apple hardware and will subscribe on the web to avoid the App Store markup; Instapaper Premium if you are on Android or already subscribed.
• You want an e-reader in the loop → Instapaper for scheduled send-to-Kindle, Wallabag for ePub export to anything else.
• You refuse to pay and you refuse to self-host → Raindrop.io. Unlimited bookmarks, collections and devices, free, on ten platforms.
• You want to own the software and pay almost nothing → Wallabag at €11/year hosted, or free self-hosted.
• You want AI and total data ownership → Karakeep, self-hosted, with a local Ollama model if privacy is the point.
• Your links keep dying on you → Linkwarden. Three archive formats plus Wayback Machine, automatically, on every save.
• You need to publish a list to a class, a club or a colleague → Marqly (public boards, free, view-only, no signup to view) or Linkwarden (public collections). Instapaper, Readwise Reader and Matter cannot do this at all.
• You want the least app possible and the longest track record → Stay on Instapaper. Genuinely. It is cheaper than four of these seven over five years and it has outlived all of them.

If the retrieval answer is the one you landed on, the whole field is ranked in our AI bookmark manager guide for 2027, and the wider read-it-later category is in our read-it-later apps roundup for 2027. If you would rather see the free options only, the free bookmark managers guide for 2027 ranks nine of them with every limit named.

How to move your Instapaper library

Instapaper’s export is available on the free plan, which is better than most tools in this category manage. The steps:

1. Export on the web, not the app. In Instapaper’s account settings, export your saved articles. You get a choice of HTML or CSV. The HTML is a standard bookmarks file that most read-it-later and bookmarking apps import directly; the CSV is easier to inspect or clean up first.
2. Get your highlights out separately, before anything else. The export carries your links — not your annotations. If a decade of highlighting is the real library, connect Readwise and let it sync your Instapaper highlights and notes out first. This is the step people skip and then regret.
3. Check what you are actually holding. Open the file. If you want to inspect or de-duplicate it before committing it anywhere, our Bookmark File Viewer and Duplicate Bookmark Finder both run in the browser from the free tools directory.
4. Import into the destination. Marqly’s Migration Center takes browser bookmark HTML, Raindrop HTML collections, generic CSV, and the list.csv from inside a Pocket export archive. Wallabag imports Instapaper directly, which is the shortest path on this list. Karakeep’s importer list covers Instapaper too.
5. Expect to lose some things, and know which. Across every importer we have tested, the fields that survive are title, URL, description or note, tags, and folder structure. The fields that routinely do not are original save dates, highlights, article text, and read status. Folder trees get flattened — in Marqly’s case to at most two board levels. Our import-fidelity study documents the measured version of this rather than the promised version.
6. Keep the export file. Not for a week — indefinitely. It is the only copy of your Instapaper library that does not depend on Instapaper, and if the permanent archive matters to you, that archive is the one thing Premium cannot hand back when you leave.
7. Do not delete your Instapaper account on day one. Run the new tool alongside it for a fortnight. Most migration regret is about workflow rather than data, and workflow is only visible after two weeks of real use.

Tool-specific walkthroughs: the Instapaper to Marqly guide, the Pocket migration guide if you are rescuing an old archive rather than a live one, and our explainer on what is actually inside a Pocket export file.

Why retrieval beats minimalism in 2027

Instapaper’s minimalism is not a mistake, and it is not going away. It is a coherent bet that the valuable part of a saved article is the reading of it, and that everything else — tags, recommendations, intelligence layers, social surfaces — is friction between you and the text. For a person who saves an article and reads it that week, that bet pays. It has paid for eighteen years.

The bet stops paying when the pile outgrows the reading. At some point the library stops being a queue and becomes an archive, and an archive you cannot search is not an archive — it is storage. This is the same failure mode that turned Pocket into a graveyard for millions of people who only discovered the problem when Mozilla gave them a four-month window to export. Instapaper’s free tier accelerates it, because a free account cannot full-text search its own saves; the pile grows and the retrieval stays at zero until you pay.

That is why this list is ranked the way it is, and why the top row is the one with the retrieval layer. Every other axis here — typography, audio, e-ink delivery, ownership, archiving — has an excellent answer somewhere in the seven. Only one axis has an answer that most of these tools do not attempt at all: type what you remember about a thing you saved eight months ago and get it back.

Marqly was built around that sentence. The free plan stores up to 100 bookmarks with keyword search across the whole library, needs no card, and includes boards, public board sharing and six-colour highlighting on every plan. Pro at $72/year or $9/month adds semantic search, automatic tagging on every save, AI summaries, Ask with cited answers, and YouTube summaries, transcripts and chat — with a standing $49 first-year offer, which is 31% off the $72 list price, and $48 in the first year for verified students. Import your Instapaper export, your browser bookmarks or the list.csv from a Pocket archive through the Migration Center, then try describing something you saved and cannot name.

Get started free →

---

Prices, plan limits, platform lists and shutdown dates verified against our dated competitor fact sheets in October 2026; the verification date for each tool is in the comparison table above, and the Pocket timeline was checked against Mozilla’s shutdown notice. Changes from the 2026 edition: Instapaper’s table price corrected from “$3/mo” to $5.99/month or $59.99/year; Readwise Reader corrected from a stale “$12/mo” and a false “semantic search ✅” to $9.99/month billed annually ($119.88/year) with keyword-only search; Raindrop corrected from “$3/mo” to $28/year or $3.25/month with its unlimited free tier named; Marqly’s pricing specified as $72/year or $9/month with the 100-bookmark free plan and Pro-only AI stated; Pocket and Omnivore removed as live options and replaced by a durability ranking; Wallabag, Karakeep and Linkwarden added; Matter’s missing Android app, missing public API and App Store price markup disclosed; the CTA changed from a Pocket-specific landing page to the standard signup; and every CTA verb on the page normalised to “get started free.” We publish no numeric ratings for any product. We will update this guide as plans change.