Lovable
Fast at the three-pane reader UI and Supabase tables.
Reading · updated 2026-11-03
Can you vibe code an RSS reader like Feedly?
Medium: Yes, with care Yes for a personal reader. Fetching and parsing feeds is easy until you meet messy real-world feeds, safe HTML rendering and polling at scale.
4–10 h
MVP build time
$0–$40/mo
Your running cost
$6/mo
Feedly Pro
up to $72
Saved per year
The honest verdict
A personal RSS reader is one of the most rewarding vibe coding projects: subscribe to feeds, fetch them on a schedule and read articles in a clean layout. The tricky parts are robustness and safety. Real feeds are malformed, use odd encodings and change URLs, and their HTML must be sanitised before you render it. Fetching arbitrary URLs from your server also needs SSRF protection. Excellent indie readers exist if you would rather not maintain one.
Feedly is rSS and news reader that collects articles from sites, newsletters and feeds, with AI filtering on paid plans. This page is about building your own version for personal or internal use, or to learn how this kind of product works, not about copying Feedly to compete with it. How we grade.
What's easy and what's hard
Easy with AI
- Subscribing to feeds and listing articles
- Read and unread state, starring and folders
- A clean reading view
- OPML import and export
- Keyboard shortcuts
Hard (where AI won't save you)
- Parsing malformed RSS, Atom and JSON Feed with odd encodings
- Sanitising article HTML so feeds cannot run scripts in your reader
- Polling politely with ETag, Last-Modified and backoff
- Full-text extraction for feeds that only publish summaries
- Native mobile apps with offline sync
Feedly feature checklist
Difficulty is for a solo builder using an AI coding tool. Skip means not worth building yourself.
| Feature | Difficulty | Why |
|---|---|---|
| Feed subscription and discovery | Easy | Detect feed links from a site's HTML head. |
| Scheduled fetching | Medium | Cron worker with conditional GET (ETag, Last-Modified) and per-feed backoff. |
| Robust parsing | Medium | Use a mature parser library and handle RSS, Atom and JSON Feed. |
| Read, star and folders | Easy | Per-user item state table. |
| Safe article rendering | Medium | Strict HTML sanitiser, proxied images, no scripts or iframes. |
| OPML import and export | Easy | Standard XML format other readers use. |
| Full-text extraction | Medium | Readability-style extraction for summary-only feeds. |
| Newsletters to feed | Medium | A unique inbound email address per user, parsed into items. |
| AI summaries and filters | Medium | LLM per article costs add up; batch and cache. |
Build time and running costs
Your own version vs Feedly, at personal or small-team scale.
Build time
- Working MVP for yourself
- 4–10 h
- Polished, relied on daily
- 4–10 days
- MVP time at $50/h equals
- 58 months of Feedly
Feedly price (not independently verified)
Pro: $6/mo billed annually
Pro is commonly listed at about $6 per month billed yearly or $6.99 monthly; Pro+ with Feedly AI is about $8.25 per month billed yearly. The free plan follows up to 100 sources. Feedly's pricing page did not load for us.
Source: feedly.com/i/pro. Prices change; check before deciding.
| Running cost | Per month | Notes |
|---|---|---|
| Scheduled fetch worker | $0–$5 | Cloudflare Workers cron or a small VPS covers hundreds of feeds. |
| Database | $0–$25 | Supabase free tier, pruning old items. |
| LLM summaries | $0–$10 | Optional; depends on how many articles you summarise. |
| Total | $0–$40/mo | Saves up to $72 a year vs Feedly Pro; at the top of the range it costs $408 a year more. |
Recommended vibe coding tools
Claude Code
Good at the fetcher: conditional requests, parsing edge cases, SSRF protection and tests with real feeds.
Cursor
Useful for iterating on the sanitiser and reading view inside your repo.
Starter prompt
Paste it into your tool, then ask for a plan before any code. Edit the specifics to match what you need.
prompt.txt
Build a personal RSS reader for my own use. Stack: Next.js, Supabase (auth, Postgres), and a scheduled fetch worker (Cloudflare Workers cron or a Node cron job). Data model: feeds (id, url, site_url, title, etag, last_modified, next_fetch_at, error_count, last_error), subscriptions (user_id, feed_id, folder, custom_title), items (id, feed_id, guid, url, title, author, published_at, content_html_sanitized, summary), item_states (user_id, item_id, read_at, starred). Requirements: - Add a feed by pasting a site or feed URL; discover feed links from the page head. - Fetch due feeds every 15 minutes with conditional GET using ETag and Last-Modified; back off on errors and 429s. - Parse RSS, Atom and JSON Feed with a mature library; dedupe by guid or URL. - Three-pane UI: folders and feeds, item list, reading view; j and k shortcuts; mark all read. - OPML import and export. Security: - Only fetch http and https; resolve DNS and block private, loopback and metadata IPs to prevent SSRF, including after redirects; cap response size and time. - Sanitise all feed HTML with a strict allowlist; strip scripts, iframes, forms and event handlers; proxy images. - Row Level Security on subscriptions and item states. - Escape feed titles everywhere.
Security pitfalls to check
AI tools often get these wrong by default. Check each one before anyone else uses the app.
- 01SSRF: a feed fetcher requests any URL a user pastes, so block private IP ranges, metadata endpoints and unsafe redirects.
- 02Feed content is third-party HTML; render it only after strict sanitisation or it becomes stored XSS in your reader.
- 03Huge or slow feeds can exhaust your worker; set size limits, timeouts and per-feed backoff.
- 04Hotlinked images reveal your IP and reading habits to publishers; proxy them if privacy matters.
- 05Shared feed rows with per-user state need Row Level Security so users cannot see others' subscriptions.
Indie makers who built a real reading business
They went far past the prototype: support, payments, reliability and years of iteration. Their profiles show what that took.
- BU
- BS
Brent Simmons built NetNewsWire
Free, open-source RSS reader for Mac, iPhone and iPad.
- SR
Silvio Rizzi built Reeder
Unified timeline reader for RSS and social feeds on Apple platforms.
Craig Hockenberry built Tapestry
Timeline app combining RSS, Mastodon, Bluesky and more.
Or just use these indie-made tools instead
If you need Feedly's job done rather than a project, these are built and supported by independent makers.
NetNewsWire
netnewswire.com
Free, open-source RSS reader for Mac, iPhone and iPad.
by Brent Simmons
Alcove
alcove.news
A calm RSS reader built in public as a personal side project.
by John O'Nolan
Feedbin
feedbin.com
Paid reader for RSS feeds, email newsletters and other web content.
by Ben Ubois
R
Reeder
reederapp.com
Unified timeline reader for RSS, YouTube, Reddit, Mastodon, Bluesky and podcasts on iPhone, iPad and Mac.
by Silvio Rizzi
Tapestry
usetapestry.com
Unified timeline app that combines Mastodon, Bluesky, RSS, YouTube, podcasts and other feeds in one place.
by Craig Hockenberry
What building it teaches you
You learn scheduled fetching, HTTP caching headers, XML parsing, HTML sanitisation and SSRF defence.
Can you vibe code these too?
- Acuity SchedulingMedium
- CalendlyMedium
- DocSendMedium
- EventbriteMedium
- FreshBooksMedium
- Google AnalyticsMedium
Vibe coding Feedly: FAQ
Can you vibe code an RSS reader like Feedly?
Medium. Yes for a personal reader. Fetching and parsing feeds is easy until you meet messy real-world feeds, safe HTML rendering and polling at scale.
How long does it take to vibe code an RSS reader like Feedly?
A working personal or internal MVP takes about 4–10 h with a tool like Lovable. A polished version you would rely on every day takes about 4–10 days, mostly spent on parsing malformed rss, atom and json feed with odd encodings.
How much does it cost to run your own Feedly alternative?
Expect $0–$40/mo at personal or small-team scale (scheduled fetch worker, database, llm summaries). Feedly Pro costs $6/mo (not independently verified), so the most you would save is $72 a year, before counting your time.
Is building your own RSS reader a good vibe coding project?
Yes, it is one of the best. A useful reader takes a few hours, and making it robust teaches you scheduling, HTTP caching, parsing and safe HTML rendering.
What is the main security risk in an RSS reader?
Two things: rendering untrusted feed HTML without sanitising it, which allows XSS, and fetching arbitrary URLs from your server, which allows SSRF unless you block internal addresses.
What are good indie alternatives to Feedly?
NetNewsWire is free and open source, and Feedbin and Reeder are long-running paid readers made by independent developers.
Is it legal to build your own Feedly clone?
Building your own tool that solves the same problem is common and generally fine. Copying Feedly's code, design assets, brand name or content is not. Use your own name and design, respect their terms of service, and get proper advice if you plan to sell it. This page is not legal advice.