How sip compares to other readers
—「Taste it. Savor it.」
The feed-reader ecosystem is crowded. sip isn't trying to compete on "aggregating subscriptions and displaying articles" — existing readers already do that well. What sip wants is what they generally don't do and won't do: making information trustworthy, traceable, and usable by AI.
Below we compare against the common three categories. The high-level conclusion first, then the details.
Positioning in one line
| Category | Examples | Core experience | What sip offers differently |
|---|---|---|---|
| Self-hosted web readers | FreshRSS, Tiny Tiny RSS, Miniflux | Deploy it yourself, read feeds in a browser | sip is a local single file, plus version tracking / AI whitelist / one bowl a day |
| Cloud aggregation services | Feedly, Inoreader, The Old Reader, NewsBlur | Hosted aggregation, sync across devices | sip keeps data in your hands, no account, no upload; AI retrieval reads a whitelist only |
| Desktop / mobile clients | Reeder, NetNewsWire, Follow | A beautiful local reading UI | sip is a CLI + TUI built for scripts / agents, not just a UI for humans |
Self-hosted: FreshRSS / Tiny Tiny RSS / Miniflux
They all help you "aggregate subscriptions and read in a browser." That space is mature; sip doesn't reinvent it.
What sip adds (no feature-talk here — every command really runs):
| sip capability | Can they? | What it looks like |
|---|---|---|
| Version tracking + diff | ❌ they only show "new article" | sip --versions 42 / sip --diff 42 — edits, corrections, retractions all recorded |
| AI reads a whitelist | ❌ | sip --search "topic" --json — semantic search / summaries only from your sources |
| One bowl a day | ❌ they give you an endless feed | sip --today — 5 rule-based picks daily, anti-scroll, anti-bubble |
| Agent-facing CLI | ⚠️ APIs exist but are heavy, human-oriented | Unified JSON + structured exit codes + sip-rss skill, drop straight into a bot |
| Deployment | Server + database + web | Local single-file exe, all your data in readwithhotsoup/ |
Cloud: Feedly / Inoreader / The Old Reader / NewsBlur
Their strength is "hosting + multi-device sync + recommendation algorithms."
sip's differences:
- Data ownership: they store reading history and feed lists in the cloud; sip keeps everything local in
readwithhotsoup/(SQLite + file cache), no account needed, reading history never uploaded. - Transparent decisions: cloud "recommend this to you" is a black box; sip's filter rules are simply the feed list you maintain, and every pick can say "why" (Transparent Decisions).
- AI capability: most clouds don't have "let AI read only your whitelist" — either no AI, or the AI still cites garbage sources from the whole web.
Clients: Reeder / NetNewsWire / Follow
They perfect the reading UI. sip's TUI and immersive reading pay respect to them, but the positioning differs:
- They are UIs for humans; sip is a tool for humans + AI — the same data, humans read in the TUI, agents query with
--search --json. - Clients generally don't solve "articles get edited" or "AI cites trustworthy sources" — which is exactly sip's core.
The common blind spot: three things almost nobody does
Regardless of category, most readers get stuck in the same place — "aggregate + display", not "trust + use":
| Blind spot | Traditional readers | sip |
|---|---|---|
| Author edited the article, what changed | You keep seeing the old version | --diff shows the evolution (Smart Archiving) |
| Want AI to research without citing garbage | No answer | AI only retrieves your subscribed sources (AI Friendly) |
| Information overload, endless scrolling | Give you more | One bowl a day, read and close (Today) |
When you should NOT use sip
To be honest, sip isn't right for these:
- You want a web-based reader accessible from a phone browser → choose a self-hosted reader like FreshRSS / Miniflux.
- You want multiple users online, sharing one subscription and account system → self-hosted or cloud.
- You just want 200 feeds in one place and don't care about "article edits" or "AI citations" → any mainstream reader is enough.
- You already have a mature deployment and migration costs are high → no need to switch.
Migrating over?
Mainstream readers all support OPML, and so does sip:
sip --import-opml your-export.xml # FreshRSS / Feedly / Inoreader / TT-RSS all export OPML
sip -l # confirm the feeds are inRunning both in parallel for a week works fine: the old reader for quick browsing, sip for deep reading and AI. Keep the one with good reason, remove the redundant one.
Next steps
- 5-Minute Quick Start — from download to first value
- Practical Scenarios — 6 real workflows
- Features Overview — the three capability modules