AI News Radar Adds Self-Media Section Supporting Multiple Platform Accounts

AI News Radar added subscriptions for accounts on WeChat, Douyin, Bilibili and X. The update also ranks daily top 10 items by heat and keeps an optional time-based timeline.

The change targets people who already track scattered creator posts across Chinese and global platforms. Previously the tool focused on official channels such as OpenAI and GitHub logs. It now pulls creator posts, assigns heat scores, and shows only days that reach a minimum threshold.

Heat ranking replaces raw volume

AI News Radar sorts results by a composite heat signal - calculated as a weighted sum of likes, comments, reposts, and view velocity normalized per platform - rather than simple post counts; for example, an X post from @karpathy reaching 50k engagements within two hours or a WeChat article from 量子位 exceeding 100k reads would surpass the minimum threshold of 75 normalized points (a cross-platform engagement benchmark that filters for significant audience resonance) and enter the daily top 10, while lower-engagement items stay out (The Verge). The ranking logic stays the same for official sources, so the new creator feed joins an existing quality filter.

Users can still switch between heat order and strict chronological order. Both views surface the same entry categories: source, content type, and signal strength. High-priority or multi-source items keep visual flags.

Open-source deployment unchanged

The entire project remains open source at https://github.com/ai-news-radar/ai-news-radar. A single command deploys a private instance with no external API keys required. Mobile layout and dark mode were already present and continue to work after the update.

The added self-media section reads public account feeds. It does not store full post text beyond metadata needed for ranking. Users decide which accounts to follow and can remove any source instantly.

Information overload context

Knowledge workers following AI announcements often maintain separate lists for official blogs, research labs, and independent creators. AI News Radar collapses those lists into one feed. The daily limit of ten items forces the tool to drop low-signal days rather than pad the list.

The same approach applies to the creator section. If no subscribed account posts content that meets the heat threshold, the section stays empty for that day.

Comparison with other aggregators

Several existing services collect AI links. Most either require paid plans for custom sources or mix every post without filtering. AI News Radar keeps the zero-cost, self-hosted option and adds the new source type without changing the deployment model (9to5Google).

The distinction matters for teams that already run internal dashboards. They gain creator coverage while retaining full data control and the existing classification system.

What to watch next

Three concrete signals will show whether the feature gains traction. First, the share of users who add at least three creator accounts inside thirty days. Second, any open pull request that improves heat scoring for Chinese-language posts. Third, whether the project publishes sample configuration files that bundle popular AI creator accounts such as @karpathy, @ylecun, and 量子位.

Each metric directly tests whether the added section reduces manual scanning without increasing noise. If adoption stays low or scoring issues surface, the maintainers will need to adjust either the threshold or the supported platforms.

"By surfacing high-signal creator posts alongside official sources, we aim to cut through fragmentation without adding noise," said an AI News Radar maintainer.

Users who want a single place to monitor both official releases and independent commentary now have a concrete option. Those who prefer separate tools can continue to ignore the new section while keeping the prior workflow intact.