Features – Readfine

Feeds & subscriptions

Get content in, from anywhere.

RSS & Atom feeds: Subscribe to any standard feed URL.
Web-scraping feeds: Follow sites that have no feed. Point Readfine at a listing page and a CSS selector that matches the article links; with an AI key it can suggest the selector for you.
Authenticated feeds: Subscribe to feeds behind HTTP Basic Auth with per-feed credentials, stored encrypted.
Folders: Group feeds into folders for organizing and for scoping filters.
Scheduled background fetching: New articles are pulled automatically on a schedule, with adaptive per-feed intervals that back off quiet feeds and check busy ones more often.
OPML import & export: Move your subscriptions in and out, including files compatible with Tiny Tiny RSS.

A clean reader that adapts to your screen.

Readable extraction: The full article text is pulled into the reader (trafilatura, with readability-lxml as a fallback), without the cookie banners and sidebars of the original page.
Adaptive layout: Choose a 2- or 3-panel view per screen size, with breakpoints you set yourself.
Dedicated mobile layout: A real small-screen layout with a collapsible sidebar and inline or full-screen article view, not a squeezed-down desktop.
Search: Full-text search across your articles, scoped by feed, folder, status, or label. Open it from anywhere with the / shortcut.
Dark mode: Light, dark, or follow your system preference.
Reading typography: Choose the font family and text size for the reading view.
Number & date format: Pick how numbers and dates are written (decimal separator and date order) independently of the interface language, so English can pair with 1 234,56 and 25.06.2026.
List density: Compact, comfortable, or summary, set separately for desktop and mobile.
Mark read on scroll: Optionally mark articles read as they scroll past.
Article states: Mark articles read, star favourites, and archive what you want out of the way.
Labels: Your own colour-coded tags for sorting articles by hand.

Trim the stream down to what matters.

Filters: Match on title, content, author, URL, or AI score, then automatically add a label, mark read, star, or archive.
Regex and AND/OR: Combine conditions with AND/OR and match with regular expressions (Python syntax) when plain "contains" is not enough.
Feed & folder scoping: Restrict any filter to specific feeds or folders.
Retroactive apply: Run a filter over existing articles, not just new ones.
Relevance scoring: With an AI key, each incoming article is scored against your interest profile, so you can surface or hide articles by score (a filter can mark anything below a threshold as read).

Optional AI on your own provider key.

Summaries: A one-tap summary of any article.
Chat over articles: Ask questions about an article and get answers grounded in its text.
Interest profile: Describe what you care about, or have the AI draft it from what you read; it drives scoring and digests.
Catch me up: An on-demand digest of what happened in your feeds over a period and scope you choose.
Scheduled briefings: The same digest on a schedule, delivered to your inbox.
Your choice of provider: Anthropic, OpenAI, or Google Gemini. Keys are per-user, stored encrypted, and used only for your own requests.
Custom prompts & models: Set a fast model for scoring and a quality model for summaries, chat, and briefings, each on the provider you pick. The summary and context prompts are editable, and each briefing can carry its own prompt.

See how you actually read.

Reading stats: Your read rate, reading streak, average dwell time, and your single most active day and hour.
Scoring calibration: Check how well AI relevance scores match what you actually read, and surface "missed gems" the AI rated highly but you never opened.
AI usage & cost: An estimate of your AI spend per provider over time.

Multi-user, with the operator controls to match.

Per-user accounts: Every account has its own feeds, filters, labels, and preferences.
Admin panel: Manage users and instance-wide settings.
Email (SMTP): Address verification, password reset, and briefing delivery.
API tokens: A JSON API authenticated with JWT tokens, for scripts and integrations.
Retention & purge: Tiered retention rules that purge old articles automatically.

Self-hosted and fully open source.

Self-hosted: Runs on your own server with Docker; your data stays with you.
Open source: AGPL-3.0 licensed, so you can read and modify every line.
Encrypted key storage: AI keys and other secrets are stored encrypted at rest.
SSRF protection: Feed and scrape fetches are guarded against server-side request forgery.
Hosted option: Prefer not to self-host? Use the hosted instance at readfine.app.