Cycling through all six. Tap any point to stop.

Measured on six things

Bring RSS, Telegram, X and YouTube into a feed that learns from what you approve, reject and explain. Repeating coverage of the same event goes into one story with original links. Built by a newsletter editor. From $9/mo.

Security

HTTPS enforced and secure cookies set, but no HSTS, CSP, frame protection, Referrer-Policy, Permissions-Policy or DMARC — 4/13 header checklist points. Telegram access is claimed read-only with encrypted credentials; no SOC 2/ISO/pen-test published.

Screenshots

1 / 3

Radaric

About this app

Bring RSS, Telegram, X and YouTube into a feed that learns from what you approve, reject and explain. Repeating coverage of the same event goes into one story with original links. I edit Zhurnalus, a weekly design digest with 500+ issues since 2016, and Radaric started as the tool I built for my own curation. The part that matters most is teaching the filter. "Design" is a starting point, but it does not capture an editor's taste: a case study, a minor product update and corporate PR can all match the same keywords. You describe what matters in plain language, then approve or reject posts, adding a reason when it helps. That feedback gives the AI context, and you can see its reasoning and revisit what it filtered out. RSS, Telegram including private channels, and YouTube live in one inbox, with X on Pro. When several sources cover the same event, the posts become one story with the original links attached. Lite is $9/month, Pro $59/month. 30-day money-back guarantee.

Key features

1Taste-learning AI filter that improves from approvals, rejections and explanations
2Cross-source story deduplication across RSS, Telegram (private), X, YouTube, Substack, Bluesky, Mastodon, Product Hunt, GitHub and arXiv
3YouTube transcript reading and private Telegram channel support
4OPML import from Feedly/Inoreader plus one-click source adds

Target customer profile

Drawn from the product itself, not from a survey.

Demographic

Newsletter writers, editors, and independent publishers

Pain points

Drowning in hundreds of feeds; repeating triage across 15 sources covering the same event

Primary needs

Faster discovery of genuinely relevant stories and automatic deduplication

Demographic

Researchers, analysts, and product/design professionals

Pain points

Scrolling endless social and RSS feeds; losing hours to manual curation

Primary needs

A personalized, low-noise feed that learns their niche and filters fast

Demographic

Content creators and social media managers

Pain points

Juggling Telegram, X, YouTube, and Substack; missing cross-platform trends

Primary needs

One unified inbox with multi-source coverage and deduplicated news

Demographic

Power users of Feedly, Inoreader, or Telegram-first readers

Pain points

One-size-fits-all AI filters and no support for private Telegram or YouTube transcripts

Primary needs

A reader whose AI adapts to their personal taste and covers more source types

Scout Review

Written by AI from measured evidence, scored out of 100.

Radaric is a taste-learning feed reader built by an editor who clearly uses it daily, and that shows in the copy, the onboarding flow, and the cross-source story deduplication that actually merges a Russian Telegram channel, an English X thread and a Portuguese RSS feed into one story. The wedge is real: Feedly's AI is the same for everyone, Inoreader has no taste-learning, and neither handles private Telegram or YouTube transcripts. The execution has rough edges — mobile LCP at 6.6s, a security header posture of 4/13, a comparison table with no headers, and no free tier — but none of them are fatal. The pricing is the biggest question: $59/month for X parsing against Feedly Pro+ at $18 is a steep ask for an early-stage product.

Measure by measure

Usability

Four-step onboarding is spelled out clearly, OPML import from Feedly/Inoreader is offered, and one-click adds cover Substack, Bluesky, Mastodon, Product Hunt, GitHub and arXiv. No free tier, but a 30-day money-back guarantee softens the trial.

Design

Confident editorial landing page with numbered steps, a Feedly/Perplexity comparison table, and a named founder with a 500-issue track record. Copy is specific and on-brand for editors.

Speed

Lighthouse desktop 93/100 but mobile 71/100 with a 6.6s LCP — well past the 2.5s 'good' threshold. CLS 0 and TBT 51ms are clean; the mobile hero is the bottleneck.

Security

Accessibility

Lighthouse 91/100 with three real failures: insufficient color contrast, viewport blocking user zoom (user-scalable=no), and a large table with headerless <td> cells. No WCAG statement published.

Growth

Real problem (cross-source triage + dedup), specific target (newsletter editors), and a genuine wedge — taste-learning filter plus private Telegram and YouTube transcripts that Feedly/Inoreader structurally don't do. Early stage (domain 0.6 years), so headroom is real but unproven.

Improvement areas

Mobile LCP of 6.6s fails Core Web Vitals — compress hero assets and preload fonts
Security headers at 4/13 — add HSTS, CSP, frame protection, Referrer-Policy, Permissions-Policy and DMARC
Accessibility: add <th> headers to the comparison table, remove user-scalable=no, fix contrast
Pricing gap between $9 Lite and $59 Pro with no middle tier and no free trial

Design & Usability

Radaric's site reads like it was written by someone who actually uses the product — because it was. The numbered 01-04 flow, the Feedly/Perplexity comparison table, and the named founder with 500+ Zhurnalus issues give it more credibility than most indie SaaS landing pages. Onboarding is spelled out clearly: connect feeds, AI filters, AI learns, read. OPML import from Feedly/Inoreader is a smart migration path, and one-click adds for Substack, Bluesky, Mastodon, Product Hunt, GitHub and arXiv broaden the surface well beyond RSS. The main usability friction is the absence of a free tier — the 30-day money-back guarantee is a reasonable substitute, but it asks for a credit card before the user has seen the filter learn anything.

Speed & Security

Desktop Lighthouse lands at 93/100, but mobile is 71/100 with a 6.6-second LCP — the hero is heavy enough to fail Core Web Vitals on phones, which is where most feed-reading happens. CLS 0 and TBT 51ms show the underlying build is clean; this is a fixable asset problem, not an architecture one. Security is the weaker story: HTTPS and secure cookies are in place, but the header probe found no HSTS, CSP, frame protection, Referrer-Policy, Permissions-Policy or DMARC — 4/13 checklist points. For a product that asks users to connect private Telegram channels and X accounts, that's a thin posture. The site claims read-only Telegram access and encrypted credential storage, which is the right claim, but no SOC 2, ISO, pen-test or bug bounty is published to back it up.

Accessibility & Growth

Accessibility scores 91/100 on Lighthouse, but the three failing audits matter: insufficient color contrast, a viewport that blocks pinch-zoom on mobile, and a large comparison table with headerless <td> cells. The last one is ironic for a product aimed at editors — a screen reader user hits that table and gets a wall of unlabeled cells. No WCAG statement is published. Growth is the more interesting dimension: the problem is real (cross-source triage and deduplication across RSS, Telegram, X, YouTube), the target is specific (newsletter writers and independent publishers), and the wedge — a filter that learns from approvals and rejections, plus private Telegram and YouTube transcripts — is something Feedly and Inoreader structurally don't do. The category has headroom, but the $59/month Pro tier against Feedly's $18 and Inoreader's lower pricing is a hard sell for a two-person shop.

Conclusion

If you're an editor or newsletter writer drowning in 15 sources covering the same event, Radaric is worth the 30-day money-back trial — the dedup and taste-learning are genuinely differentiated. If you're a casual RSS user, Feedly or Inoreader will do the job for less. The product needs to fix its mobile LCP, ship a real security header set, and either justify the $59 Pro tier or bring it down before it can compete on price.

Suggestions for improvement

Fix the mobile LCP (6.6s) — compress the hero image and preload the primary font to get under 2.5s.
Add HSTS, CSP, X-Content-Type-Options, Referrer-Policy, Permissions-Policy and DMARC — 4/13 is thin for a product handling private Telegram credentials.
Add <th> headers to the Feedly/Perplexity comparison table and remove user-scalable=no from the viewport meta.
Reconsider the $59 Pro tier — Feedly Pro+ is $18 and Inoreader is cheaper; either justify the gap with more Pro-only sources or lower the price.

How we score

Alternatives comparison

Named competitors, point by point. Nobody paid to appear here or to be left out.

	Radaric	Feedly	Inoreader	Research assistant tools (e.g., Perplexity)
Price (entry tier)	$9/month (Lite)	~$18/month (Pro+)	Free tier; paid from ~$9/month	$2,000+/month (enterprise)
Taste-learning AI filter	Yes — learns from approvals, rejections and explanations	No — same AI for all users	No — rule-based automation only	No — general research assistant
Private Telegram channel support	Yes, automated	No	No	Manual only
Cross-source story deduplication	Automatic grouping with original links	No	No	Manual
YouTube transcript reading	Yes	No	No	No
Free tier / trial	No free tier; 30-day money-back guarantee	Free tier available	Free tier available	Free tier available

Feedly

Popular RSS reader with AI filtering (Pro+ around $18/month), but its AI is the same for every user and it focuses on RSS rather than Telegram, YouTube transcripts, or private channels.

Inoreader

Feature-rich RSS reader with automation rules and OPML support, but lacks taste-learning AI and automatic cross-source story deduplication.

Research assistant tools (e.g., Perplexity)

Enterprise-grade AI research assistants can cost $2,000+/month and require manual setup, versus Radaric's from-$9/month approach focused on personalized feed curation.

Comparing options? See Radaric alternatives, scored side by side

View all Productivity