FlowReader Pro

A reading system that keeps working

Pro adds collection, signals, briefings, optional sync and archive-grounded AI. The local reader underneath stays exactly as it is.

See pricingCompare with free

01

Collect, track and brief in one Flow

Describe the outcome first. FlowReader runs it once, shows what happened, and only then offers to repeat it daily, weekly, or when new items arrive. One Flow can work from explicit watchlists, your own archive and open-web discovery, and can combine collection, tracked signals and cited briefings.

Flow lifecycle

  1. 01

    Ask

    Describe what you want in ordinary language.

  2. 02

    Run

    FlowReader plans it and runs it once, with visible steps.

  3. 03

    Read

    Inspect the collected items, signal or cited briefing.

  4. 04

    Repeat

    Only schedule the result after it proved useful.

Flow outputs

Collect

AI optional

Watch explicit pages and feeds, then archive everything new or only what matches.

Archived items

Signal

AI optional

Follow a price, sentiment, mentions or another value over time and act when it changes.

Tracked value and alert

Brief

AI required

Research your archive, selected sources or the open web and write a grounded briefing.

Cited narrative and charts

Requests a Flow can carry

  • Watch these five research blogs, archive every new post, and brief me on Friday.

    Collect and Brief in one Flow.

  • Tell me when this price moves, and show the history in my next briefing.

    Signal feeding a Brief.

  • What is trending on Hacker News today, and what are people arguing about?

    One-shot discovery and cited research.

Nothing runs on a schedule until it has earned one

A Flow is not a configuration wizard with a save button at the end. You describe what you want in a sentence, FlowReader plans it and runs it once with its steps visible, and you read what came back. Only then are you offered the choice to repeat it — daily, weekly, or whenever new items arrive. A flow you never ask to repeat simply never runs again.

  • One sentence, planned and run in front of you. Placeholder.

  • Everything it decided, open for you to correct. Placeholder.

  • A briefing, with every citation checked against what the run fetched. Placeholder.

  • What each run did, step by step. Placeholder.

What it will not do behind your back

  • Collecting is deterministic: a Flow that watches sources and archives what matches needs no AI provider at all
  • There is no FlowReader-owned AI key. A Flow runs on your provider, and the key is encrypted on your device before it is stored
  • It tells you where it ran — on this device, or on the FlowReader server so it can finish with the app closed
  • The research depth dial estimates the model calls, tokens and cost before you spend them, and says so plainly when it cannot price a model
  • Preview a collection first: it runs the real fetch and filter and writes nothing, then tells you what it would have archived
  • A briefing that cites a source the run never fetched is rejected rather than published

02

One library, on every device

Sync is not a backup you hope never to need. It is what turns the browser extension, the desktop app, the phone and the machine at work into one library instead of four. Save an article on the laptop and it is on your phone on the train, at the paragraph you stopped on.

What travels between devices:

  • Reading positions, so you stop on the laptop and continue on the phone mid-paragraph
  • The archive itself: items, collections, tags and reading status
  • Highlights, notes and annotations
  • Reader settings, presets and custom themes
  • News sites, feed subscriptions and feed settings
  • Smart Rules and reading statistics
  • Enough cached content to reopen a saved item on a device that never downloaded it

There are three ways to do it, and they are genuinely different arrangements rather than three brands of the same thing. One uses a FlowReader service. The other two use storage you already control, and encrypt the data with a passphrase only you hold.

Local only

The default. One device, nothing leaves it. No account, no server, no network dependency for reading.

Nothing is transmitted, so there is nothing in transit to protect.

Online Sync

Pro

The browser extension, the desktop app and the phone stay in step through FlowReader’s own service, in close to real time and without you doing anything. This is the one that makes a second device feel like the same library.

Protected by account authentication, TLS in transit, and per-user server rules that make one account’s data unreadable to another. It is not end-to-end encrypted: the service can technically read what it stores. If that matters to you, use folder or Dropbox sync instead.

Dropbox

Devices sync through your own Dropbox account. Slower to converge than Online Sync, and it needs Dropbox on each machine, but no FlowReader server is ever involved.

End-to-end encrypted with a passphrase you choose and only you hold: AES-256-GCM, keys derived with PBKDF2 at 600,000 iterations. Dropbox stores a file it cannot read. Lose the passphrase and the data is unrecoverable. That is what the guarantee costs.

Local folder

Devices sync through a folder you control: an iCloud Drive or Google Drive folder, a network share, anything your own file-sync tool already handles.

The same end-to-end encryption as Dropbox sync, over whichever transport you chose. FlowReader never sees the folder or the passphrase.

What each option means for your data

03

Newsletters

Generate private alias addresses and subscribe with them. Incoming newsletters are sanitized server-side and land in your archive as clean, readable items, with no inbox involved. Aliases can be disabled at any time, and senders can be blocked.

The Newsletter surface, where private FlowReader alias addresses are managed once an account is connected.
Alias management requires a signed-in account. Shown here in its signed-out state.
Placeholder: a newsletter issue in the reader is not photographed yet.
A delivered issue, sitting in the archive as an ordinary readable item. Placeholder.

04

Smart Rules

Persistent local automations that keep the library organized without manual filing. Condition trees combine nested AND/OR groups and negation across metadata, reading behaviour, and optional AI conditions; actions add to a collection, add a tag, or mark as read, and are always additive and idempotent.

  • Nested conditions
  • AND / OR groups
  • Negation
  • Metadata conditions
  • Reading-behaviour conditions
  • Optional AI conditions
  • Tag and collection actions
  • Rule priorities
  • Stop on match
  • Live rule testing
  • Audit history

05

Knowledge graph

A local graph across everything you save: items, collections, tags, annotations, extracted entities and the topics that emerge from them. It is built on your device from the archive you already have, and it finds the clusters you never named and the bridge items connecting otherwise separate areas of your reading.

The knowledge overview, showing archive coverage filters, suggested questions and activity across saved sources.
Placeholder: emergent topics are not photographed yet.
Placeholder.

A graph over a small library is mostly noise: the structure needs volume before it means much.

06

Archive-grounded AI

Document chat and collection chat, with inline citations that jump back to the exact source passage. Tools cover archive search, opening items, reading the current page, web search and fetch behind per-origin consent, persistent memory, and the knowledge-graph toolset. Skills and sub-agents extend it; MCP servers bridge in external tools.

The AI assistant panel in FlowReader before a provider is configured, showing that the assistant runs on user-supplied credentials.
The assistant before a provider is configured.
Placeholder: the AI chat citation view is not photographed yet.
An answer with citations back to the passages it came from. Placeholder.

Routing covers Anthropic, OpenAI, Google and Hugging Face, or fully local models via Ollama and LM Studio. You supply the credentials. There is no app-owned AI key reading your data.

07

Your library, from your AI client

FlowReader runs a Model Context Protocol server, so you can authorize Claude — or any other client that speaks the protocol — to work with your library directly. Not a chat window bolted onto the app: your own assistant, searching what you have saved, filing it, highlighting a passage, saving a page you point it at, or starting a Flow and reading what it produced.

One endpoint, and an authorization you control

Connect over OAuth with PKCE. FlowReader never asks the client for your password, and you are shown the client and where it will send you back to before anything is granted.

https://mcp.flowreader.io/mcp

You approve each permission separately, and you can approve less than was asked

The consent screen lists one switch per permission the client requested. Reading is the floor and stays on; everything else is yours to refuse, which is how you give a client read-only access to your library. Fetching public pages is held apart from writing on purpose — letting something change your archive should not silently include letting it go out to the internet on your behalf.

  • ReadAlways on

    Read your synced library, highlights, settings, history and accepted shares.

    flowreader:data:read

  • Write

    Save and change items, collections, tags, highlights, sources and automations.

    flowreader:data:write

  • Delete

    Delete items and organisation, and only after confirming each deletion.

    flowreader:data:delete

  • Settings

    Change reading, library, feed and AI preferences, themes and presets.

    flowreader:settings:write

  • Shared archives

    Contribute to archives other people have shared with you, where they allow it.

    flowreader:shared:write

  • Fetch web pages

    Fetch public web pages you ask it to read or save. Never private addresses.

    flowreader:web:fetch

What a connected client can do

Find and read

  • Search and filter your library the way the Archive does
  • Read a saved article, your highlights and your notes
  • Follow connections in your knowledge graph to find related reading
  • Read your reading history and statistics

Organise

  • Save an article written or supplied by the assistant
  • Rename, tag, file and mark items read or unread
  • Create and rearrange collections and tags
  • Add and edit highlights and notes

Automate and report

  • Create and edit Flows, Smart Rules and Skills
  • Run a Flow now, or preview what one would collect
  • Write a report from your archive that appears alongside your Flow reports
  • Read Flow runs, reports and tracked signals

Bring things in from the web

  • Fetch a public page you point it at and read it
  • Save that page to your library
  • Check what a feed you follow is publishing right now

Change preferences

  • Change reading, library, feed and AI settings
  • Save and apply themes and presets

Delete

  • Delete saved items, collections, tags and highlights
  • Every deletion has to be confirmed in the request that makes it

Shared archives

  • Add to and edit archives shared with you, within the permissions their owner set

Where it stops

The boundary is architecture, not policy. The service reads your synced library through the same per-user rules the app does; it has no route to the device. A library you keep local, in Dropbox or in a folder is not withheld from a connected client — there is nothing there for it to reach. Provider keys and billing records are excluded outright, and credential-shaped fields are stripped in both directions.

Your synced library and accepted sharesReachable
A local-only, Dropbox or folder libraryNever
AI provider keys and billing recordsNever
Fetching public pages on your behalfSeparate consent
Deleting anythingConfirmed per request
Creating a share, or running a Smart RuleNot offered

  • Every connected client, what it may do, and one button to revoke it. Placeholder.

  • One switch per permission, approved before anything is granted. Placeholder.

Requires Pro, a verified account, and Online Sync switched on.

08

Sharing, two ways

Reading is often the start of a conversation: a reading list for a team, a folder of papers for someone new to a subject, a collection built with a colleague. There are two ways to hand that over, and they solve genuinely different problems.

Share with people

A collection or a single document, shared with named people.

You invite people by email address, one at a time or several at once, and nothing is shared until they accept from their own inbox. There is no public link to leak and no unlisted URL that turns out to be guessable. If someone was not invited, there is nothing for them to open. A shared collection stays live: new items, notes and highlights appear for everyone who has access.

  • Share a whole collection, or one document
  • Invitations go to named email addresses, and there is no public link
  • Recipients accept or decline from a shared inbox, and can be removed later
  • Three roles: owner, viewer, collaborator
  • Five separate permissions: edit notes, edit tags, edit annotations, add documents, remove documents
  • Permissions are enforced on the server, not just hidden in the interface

Pro, and an account with a verified email address.

Share as a file

A portable bundle that needs no account on either side.

A collection can also leave as a single file. The .flowcollection bundle carries the items, their tags, your annotations and your reading positions, and imports into any FlowReader, with no account, no server and no invitation. It is also the honest answer to "what if I stop paying, or you disappear": your reading is a file you already have. HTML export goes further out, producing something readable in any browser forever.

  • .flowcollection bundles carry items, tags, annotations and reading positions
  • Import needs no account on either side
  • HTML export for something readable without FlowReader at all
  • Bundles are capped at 50 MB, so very large collections export in parts

Free. No account needed.

Deciding

Is Pro worth it for you?

It depends on whether your reading has to leave the device: a second machine, a mailbox, a server that keeps watching while the app is closed. Every difference is listed line by line on the pricing page, next to what it costs.

Compare Free and Pro

Add the layer above the reader

Get the desktop app or the browser extension.

Download FlowReaderHow data is handled