Readplace is a read-it-later app built from a 10-year personal reading system. Save articles with one click using the browser extension for Firefox or Chrome, read them in a clean reader view, and get TL;DR summaries for every article.

The first 50 founding members get full access free, forever. After that, $3.99/month — includes TL;DR summaries.

What happened to Pocket and Omnivore?

Pocket was acquired by Mozilla and shut down on July 8, 2025. Omnivore was acqui-hired by ElevenLabs and shut down in November 2024. Readplace was built as a reliable alternative, with an 'Even If You Cancel' promise — your data is always exportable.

What features does Readplace have?

Readplace offers browser extensions for Firefox and Chrome, a web app for managing saved articles, a distraction-free reader view, TL;DR summaries, dark mode, and secure OAuth with PKCE. Planned features include personalised AI summaries, preference learning, Gmail integration, and highlights and notes.

What does the $3.99/month subscription pay for?

Each saved article runs through a pipeline: Mozilla Readability parses the page (free, open source); real Tesseract OCR runs locally on Lambda to extract text from multi-page scanned PDFs — pixel-level character recognition, not an LLM "reading" the image, so it never hallucinates — up to 300 pages and 500 MB per file (free, open source); and DeepSeek V3.2 writes the TL;DR, disambiguates the canonical URL when an extension capture and a link submission point at the same article, and cleans up Tesseract's OCR output for structure only (paragraphs, headings, lists) before a deterministic document-diff review rejects any token the LLM tried to add or remove, so no hallucinated words ever reach you. The $3.99/month covers the infrastructure cost and crawler maintenance. There is no ad path, no data resale, and no third-party tracking.

Does Readplace hallucinate text when extracting PDFs?

No. Readplace prioritises correctness over hallucination — no AI generated slop in your PDFs. Tesseract OCR does the recognition deterministically, character by character, on the actual pixels of the page. DeepSeek is only allowed to restore structural markup (paragraphs, headings, lists) on top of the OCR output, and a document-diff review then rejects any token the LLM tried to add or remove. The words you read are the words on the page. Other read-it-later apps that 'use AI for PDFs' typically feed the image to a multimodal LLM and accept whatever it returns, which can silently substitute, summarise, or invent text — that is the AI generated slop we refuse to ship.

Read the web. Not the slop.

Paste any article link. Get a clean reader view with a TL;DR summary — articles or PDFs, in seconds. No signup. No download. Try it with that one article you've been meaning to read.

Free. No account. Works with any article.

Or keep scrolling — the story is below

A home for articles, newsletters, essays, longreads, news, blogs, stories, posts, reports, and interviews.

Pocket is gone. Omnivore is gone. I built Readplace from a personal system I've maintained for 10 years. Save articles with one click, read them when you're ready.

Install Browser Extension View on GitHub

Firefox & Chrome

Save an article. Read it later.

Desktop

Browser Extension

What works today.

These features are shipped and working right now. Install the extension and try them.

TL;DR Summaries

Every saved article gets a TL;DR outlining the most important points. Built on the same AI that powers the reading experience.

PDF Extraction with Vision OCR

Save any PDF link. Vision OCR turns it into a clean, readable article with a TL;DR — scanned pages included.

Browser Extensions

Save any page with one click, Ctrl/Cmd+D, or right-click. The extension captures the full rendered page — picking the most complete version of the content over what a URL-only crawl would see. Available for Firefox and Chrome.

Links Import

Upload bookmarks, notes, newsletters — any text-shaped export — and Readplace pulls every URL out for you to review before saving.

Privacy First

Hosted in Sydney. Australian Privacy Act compliant. No tracking, no ads.

Same article, different versions?

When you save from the browser extension, Readplace captures the rendered page you’re looking at. When you paste a link on the homepage, Readplace fetches it server-side. DeepSeek does the canonical disambiguation so both paths converge on the same article — even when the source URL is wrapped in a tracker (utm_*, gclid).

Browser extension rendered page

Pasted link server-side fetch

DeepSeek canonical disambiguation

One article trackers stripped

The extension-captured version wins when both exist, as long as it’s the most complete version of the content. If you saved from the extension while logged in to a site, you captured the full article — saving again when logged out won’t overwrite it with the paywalled version. Most read-later apps replace the captured version with whatever the canonical URL returns; Readplace keeps the most complete one.

What's Next

Save from Mobile An iPhone app that allows you to save a link on the go.

Preference Learning Resurface previously saved articles that match what's interesting to you. You can change your preferences over time.

Highlights & Notes Highlight passages and add notes as you read to help learning usingthe science of Blocked Practice.

Why I built this

I'm Fayner Brack. You might know me as the creator of js-cookie, a JavaScript library with over 22 billion downloads per year on jsDelivr. I've been building for the web for a long time.

For the past 10 years, I've maintained a personal reading system — a pipeline of Gmail filters, DynamoDB tables, and Reddit automations that helped me save, organise, and actually read the articles I cared about. That system generated 300,000+ Reddit karma across technical communities.

When Pocket was acquired and then abandoned, and Omnivore shut down overnight, I realised the tool I needed didn't exist as a product anyone could use. So I'm turning my personal system into Readplace — built in Australia, one feature at a time.

This is a solo project. I'm building it in public, and I'd rather be honest about what works today than promise features that don't exist yet.

I am the only person working on Readplace. The code is on GitHub. The infrastructure is in Sydney under Australian privacy law. If Readplace ever shuts down, your data exports to JSON, and the codebase is forkable — you can self-host the same software the day after. I cannot promise I’ll succeed. I can promise that if I don’t, you won’t lose anything.

Readplace vs the Read-It-Later Alternatives

	Readplace	Readwise Reader	Instapaper	Raindrop.io	Karakeep	Wallabag
Price	$3.99/mo	$9.99/mo	Free / Premium	Free / $3/mo	Free (self-hosted)	Free (self-hosted)
AI Features	TL;DR summary + page content capture	Ghostreader AI	✗	✗	Auto-tagging	✗
PDF Text Extraction	300 pages / 500 MB — Tesseract OCRNo hallucination. DeepSeek restores structure only.	30 MB — pre-OCR required for scans	Premium only — no OCR for scans	100 MB free / 300 MB Pro — no OCR	50 MB default — no OCR	✗
Open Source	Source-available	✗	✗	✗	✓	✓
Offline Reading	Planned	✓	✓	Pro only	✓	✓
Platforms	Mobile + Desktop Web, Chrome, FirefoxNo native app needed.	Web, iOS, Android, Extensions	Web, iOS, Android, Kobo	Web, iOS, Android, Extensions	Web, Chrome, Firefox	Web, iOS, Android, Extensions
Public Reader (no account)	Paste any link to view or share — no signup	✗	✗	✗	✗	✗
Privacy Focus	AU hosting, no third-party tracking	US-hosted	US-hosted	—	Self-hosted	Self-hosted

What Readplace will NOT become.

Readplace will not add:

Nested folder hierarchies
Social feeds or public collections
Secret browsing-history capture
Recommendation algorithms to simulate social media

Those features grow daily active users by encouraging saving. Readplace is to encourage reading what matters to you.

And that's why I need your help!

What $3.99 a month pays for.

Every saved article runs through this pipeline:

Mozilla Readability — parses the page.
Real Tesseract OCR — up to 300 pages / 500 MB. Pixel-level recognition, not an LLM guessing at an image.
DeepSeek V3.2 — writes the TL;DR and restores OCR structure. A document-diff blocks any word it tries to add or remove.

No AI generated slop — correctness over hallucination. Tesseract reads the page character by character; the LLM only restores structure, gated by a document-diff. What you read is what was on the page.

$3.99/month covers infrastructure and crawler maintenance. No ads, no data resale, no VC bridge. This is the whole business.

Price-check the providers: DeepSeek pricing. Tesseract runs in AWS Lambda on a per-request basis.

First 50 Users

Founding Member

$0 forever

Be one of the first 50 users. Get full access free, forever. Help shape the product. $3.99/month for everyone else — includes TL;DR summaries.

Save unlimited articles
Firefox and Chrome extensions
TL;DR summaries included
PDF extraction with Tesseract OCR
All features as they ship
Direct access to the developer

Signup

Don’t miss out!

Save unlimited articles
Firefox and Chrome extensions
TL;DR summaries included
PDF extraction with Tesseract OCR
All features as they ship
Direct access to the developer

Become a Member

A Pocket alternative that respects your privacy.

Every read-it-later app that shut down left users scrambling. Readplace is built differently — starting with your right to leave.

"Even If You Cancel" Promise

Export everything, anytime. Your data is yours. Cancel and your saved articles stay available for export as JSON.

Source-available on GitHub

The whole codebase is on GitHub. If I ever shut Readplace down, you can fork the repo and self-host the same software the day after.

Hosted in Sydney, Australia

Infrastructure sits under the Australian Privacy Act. No third-party tracking, no ads, no analytics inside the app.

I'm building this in public.

Try the extension, tell me what to build next. Readplace gets better with every user who shows up and says what they need.