How to use AI Briefing

AI Briefing is an AI news curation service that brings together global company blogs, developer communities, GitHub releases, and Korean tech blogs in one place.

The flow is simple: browse the feed, search when you need something, and bookmark articles you want to read again.

We also provide RSS and llms.txt. To subscribe from elsewhere, see the RSS docs.

This is the English edition. The same service is also available in Korean. Use KO in the header to switch to the Korean page you are viewing.

Besides email and password (at least 10 characters), you can sign in with Google or GitHub (the buttons appear on the sign-in page when the operator has enabled them).

Once signed in, you can use personalization such as bookmarks, likes, read markers, and favorite sources.

Sort by latest, recently added, most viewed, most liked, most bookmarked, editor's picks first, popular this week, or popular this month.

"Latest" uses the time the original was published, while "Recently added" uses the time it was added to the site, so they mean different things.

Use the left sidebar to switch between all, editor's picks, favorites, categories, and individual sources.

On mobile, editor's picks rotate one at a time. Use the previous and next buttons to move yourself, or pause the rotation. If reduced motion is on for your device, picks change only when you move them.

On mobile, the keyboard opens only when you tap the source and filter search box. Use the square filter buttons to choose all, picks, favorites, or a category.

Each page shows 24 articles, and the page number is in the URL, so you can share a specific page as is.

The all-content feed (/all) uses the same width, sidebar, and sorting as the home feed, so moving between them feels seamless.

In the trending and board sidebars, each row has a ⭐ favorite button on the left and an external link shortcut on the right.

Tabs at the top of the sidebar ("All / Favorites") let you show only your favorites or return to the full list at once.

Use the view switch above the feed to choose between list and gallery.

Your choice is saved in the browser and applies to the home feed, all-content feed, board, search, bookmarks, and tag pages.

Open an article card to see the AI summary, the link to the original, and related articles on one screen.

The time the original was actually published is shown too, so you can tell how fresh the news is.

The share button copies the current link, and Share on Threads prefills a new Threads post with the title, a one-line summary, and the link.

View counts include visitors who are not signed in. Opening the same article many times counts as one view per visitor.

When an AI summary has images, click an image to view it full screen.

Open search with the search button at the top or with Ctrl/⌘ + K.

Type two or more characters to see a preview right away. Pick a result with ↑ ↓, or open the full list with "See all results" at the bottom.

Search covers titles, translated titles, summaries, and tags.

Enter a source name with keywords, like GeekNews: ai or Reddit: llm, to search only within that source.

When there are no results, buttons to browse categories and sources appear instead of an empty screen. If you mistyped a source name, the results page tells you that source was not found.

Choose popular this week or popular this month in the sort menu of the feed or all-content feed to see the most viewed articles of the last 7 or 30 days.

The trending page organizes GitHub Trending, Hugging Face Trending, and releases separately.

If a GitHub README or Hugging Face model card has a project image, we show it as the list thumbnail and in the introduction. Otherwise we use the platform's default image.

Tag pages group related articles by the topics that AI picked out.

The releases tab shows new versions of registered open-source projects as cards.

For monorepos, you can follow only the packages or release lines you need, such as langchain-core.

The detail page shows a version dropdown, the AI summary, and a link to the original on GitHub.

Use the ⭐ icon in the left sidebar to favorite five kinds of things: sources, release targets, Hugging Face pipelines, GitHub languages, and board categories.

Favorites gather in the favorites tab at the top of the sidebar, and the bell icon notifies you when new posts or versions appear.

Bookmarked articles are always available on the bookmarks page in settings.

Like and dislike cannot both be selected. Choosing one clears the other.

You can like or dislike board posts (blog and showcase) as well as articles, counted separately from article reactions. On the board and in the all-content feed you can react right from the list cards.

If you press a button that needs sign-in while signed out, you are taken to the sign-in page and brought back to the page you were on afterwards.

When you are signed in, opening an article marks it as read.

Articles you have read appear faded in the feed, so read and unread articles are easy to tell apart.

Articles added to the site within the last 3 days get a NEW badge, and the "Recently added" sort shows only fresh ones.

On the account page in settings, check your public nickname, sign-in method, and security status, and get a password reset email for email accounts.

My posts collects posts published under your name and your drafts and archived posts by status. Posts published from shared submissions keep their review flow, and you track their progress in My submissions.

/posts has blog posts written by the operator and showcase posts that users submitted and the operator reviewed and published.

Use the tree on the left to narrow by category (Blog / Showcase: AI tools · open source · web apps · GitHub · news & articles · other), and star categories as favorites.

Like the article feed, it supports sorting (latest, recently added, most viewed, most liked, and more), date filters, and pagination, and you can react right from the list cards.

Showcase posts get a link preview card (LinkPreviewCard) of the original at the end, so you can see at a glance what the post or project is.

On a post page, Share on Threads also prefills a new post with the title, a one-line summary, and the post link.

Board posts are written by people, so they appear in the language they were written in.

All content merges the news feed and board posts in time order. It is useful when you want to follow everything new on the site regardless of where it came from.

Favorites and editor's picks filters apply across everything, so you can gather only the sources and categories you care about or only editor's picks.

Article cards show the source favicon so you can tell sources apart at a glance.

You can bookmark and like or dislike right from the cards in the all-content feed.

If you have a link worth reading, submit it at /posts/submit with the URL and a short description. It can be your own project, or someone else's open source, tool, or news article.

Check "I made this" to be shown as the creator. Otherwise you are shown as the person who shared it.

You can optionally add your GitHub profile URL. If you do, the published post shows a GitHub icon linking to your profile.

There is a monthly submission limit, and rejected submissions do not count against it (feel free to polish and resubmit).

After submitting, track the progress in /settings/my-submissions: pending → approved → AI summarizing → awaiting review → published.

Published showcase posts show the submitter's nickname, which links to the list of posts they shared. The email appears as a mailto: link only if they agreed to make it public when submitting, and a GitHub profile, if provided, appears as an icon button.

This is filled in when the operator publishes the post, so the creator credit stays even if the submitting account is later deleted.

The bell icon at the top right shows how many notifications you have not read, and opens the latest ones when pressed. It checks for new notifications about every 5 minutes, only while the browser tab is active.

You get notified about: new notices, new posts in favorite sources or categories, new versions of favorite release targets, status changes of your shared submissions, and operator replies to your messages.

Send a message to the operator at /contact. Check your messages and replies at /settings/my-inquiries.

When the operator replies, the bell icon notifies you right away.

AI summarizes every article. Articles written in English are summarized and checked against the original in English, then translated into Korean. Articles written in Korean are summarized in Korean, then translated into English.

The article page shows a one-line summary, a detailed summary, related images, and tags.

Company blog titles are rewritten from the content, and community posts are filtered by AI to remove noise.

For Reddit, we do not follow external links in the post. The post's own title and text are treated as the article.

For sites like Hacker News where the URL itself is the article, we fetch the body from that URL and summarize it.

Articles whose body we could not get are not published to the feed.

The feed is collected automatically on a schedule set by the operator.

Collection and AI processing run separately, and only articles that finished AI processing appear in the feed.

An article that appeared in Korean first may take a little longer to show its English summary. Until then, the English edition shows the original title.

Subscribe to the latest English summaries with RSS 2.0 at /en/rss.xml. No key needed.

llms.txt is a guide file that helps AI agents quickly understand AI Briefing's main pages and latest article links.

To follow a single source in the Korean feeds, add ?source=source-slug to the RSS address.

Feature updates and changes to how we operate are listed by date in the notices.

When there is a big change, such as new sources, AI processing policy changes, or layout changes, check the notices first.

Suggest news sources you want added, or report summary errors, attribution problems, and removal requests, all in one place. If you are signed in, the operator reviews and replies, and you get the reply as a notification.

Support is entirely voluntary. Whether you support us or not, the public feed and basic reading features work the same for everyone.