ProofStack

A public sample evidence audit: RSS + AI readers

This is a public sample audit. The hypothetical client is "an indie developer building an RSS + AI reader" — a topic we did not invent: it maps to two real user submissions in our pain-point library (see growth/next-case-readwise.md). The $99 audit you buy redoes all of the data below for your product: every number re-fetched, checked line by line by a human, delivered within 48 hours, at most three audits per week.

The same iron rule as the rest of this site applies: every number carries its source and observation date; anything we cannot verify says "not disclosed"; nothing is invented.

Book your audit ($99, 48-hour delivery, pay after delivery)

• Delivery date: 2026-09-04
• Data composition: fresh fetches for this audit (2026-09-04, marked "fresh fetch") + already-verified case data from our library (marked with their library observation dates)
• Matches the four deliverables promised on the audit page: category pricing benchmark / real customer search phrases / competitor search playbook teardown / non-transferable list

1. Category pricing benchmark: the paywall distribution of the reader category

1.1 Verifiable pricing of the surviving players (all quoted from pricing pages, not paraphrased)

1.2 Benchmark numbers (with a sample-size warning)

• Annual-billed monthly price range: $5.59 – $9.99; monthly range: $9.99 – $12.99. (Data points: Readwise Lite $5.59, Inoreader Pro $7.50, Readwise Full $9.99, all re-verified by fresh fetch on 2026-09-04.)
• Median: $7.50/mo (annual billing). Honest caveat: there are only 3 verifiable data points (from 2 companies), so this median has limited statistical meaning. The reliable conclusion is the range — the consensus pricing band among survivors in this category is a $5–13/mo subscription, consistent with the suggested_pricing conclusion of the readwise case in our library (evidence level B).
• The two closest comparables for you (an RSS + AI reader): Readwise Full at $9.99/12.99 (Reader is not sold separately; it is included in Full — see the pricing FAQ at readwise.io/read) and Inoreader Pro at $7.50/9.99 (AI summaries are a Pro-tier feature). At both companies, AI features sit behind the paywall.

1.3 Free-tier survival rate: the free model has a systematic death record in this category

Background casualty list (self-compiled, first-hand, in Readwise's December 2022 public beta post): Google Reader, Readability, Readmill, Oyster, Scroll and others, with the original remark "Reading software, as a business, is brutal. RIP." (blog.readwise.io/the-next-chapter-of-reader-public-beta/).

Pricing conclusion for you: within the verifiable sample of this audit, business models built primarily on free access (ad-funded: Pocket; open source / VC-subsidized: Omnivore) died 2 out of 2; the survivors are either paid-only (Readwise) or use a paywall to lock AI and core capacity (Inoreader). If your plan is "free tier + AI fully open", you are fighting the entire historical evidence of this category.

2. Real customer search phrases: autocomplete observations (fresh fetch)

Method: on 2026-09-04 we queried the public autocomplete endpoints of Google (suggestqueries.google.com) and Bing (api.bing.com/osjson.aspx) directly with 8 seed phrases and recorded the results verbatim. Autocomplete reflects the high-frequency prefixes real users type, but it provides no search volume — we do not invent any volume figures.

2.1 Category phrases: AI is a modifier, the platform is a qualifier

• Google "rss reader ai" → rss reader ai summary, rss feed ai agent, rss aggregator ai, plus one counter-signal: rss reader without ai — a group of users explicitly searching for "an RSS reader without AI". Your AI feature page needs to answer both "what can the AI do" and "can the AI be turned off".
• Bing "rss reader ai" → the completions all slide toward platform qualifiers: rss reader in edge / android / in outlook / ios / iphone / extension. Bing users look for a reader by "where it installs".
• "best rss reader" (both engines) → best rss reader 2026 (present on both Google and Bing), best rss reader reddit, plus platform splits (android/ios/mac/windows). The year token 2026 is already in the completions — demand for "best rss reader 2026"-style pages genuinely exists.

2.2 Read-it-later phrases: Pocket is dead, but its name is still the entry point

• Google "read it later app" → read it later app pocket, read it later app like pocket, read it later app open source, read it later app reddit, read it later app obsidian.
• Bing, same phrase → read it later app pocket, matter read it later app, best read later app.
• Google "pocket alternative" → pocket alternative reddit / free / open source / self hosted / kobo, pocket alternatives 2025.

Pocket shut down in 2025 (getpocket.com/farewell), yet in the completions of 2026-09-04 it is still the default byword for the read-it-later category — a dead giant's name is the cheapest traffic entry point for the surviving players. The first line of Matter's homepage hero is "Welcome Pocket Readers" (library fetch 2026-09-02); the whole category is doing this.

2.3 Competitor-alternative phrases: what people searching "readwise alternative" want

The completions for Google "readwise alternative" (fresh fetch 2026-09-04), in order: reddit, free, open source, obsidian, for kindle, self hosted, kobo, notion, readwise reader alternative.

Reading: people searching for a Readwise alternative are mainly fleeing two of its properties — price (the free / open source / self hosted trio) and dependency (obsidian / notion / kindle / kobo, i.e. "just connect to my own toolchain"). If you build a "readwise alternative" landing page, these nine completions are its section headings.

2.4 Comparison phrases: who you will be compared against

• Google "feedly vs" → feedly vs inoreader (first result), feedly vs readwise reader, feedly vs netnewswire / reeder / flipboard.
• Google "inoreader vs" → inoreader vs feedly (first result), inoreader vs readwise, inoreader vs readwise reader / newsblur / feedbin / freshrss / netnewswire.
• Google "omnivore alternative" → omnivore alternative reddit / obsidian, omnivore free alternative — nearly two years after Omnivore shut down, searches for its alternatives are still in the completions.

Cross-checked against the keyword_profile in our library (multi-engine observations): "inoreader vs feedly" hits on 4 engines, best position 1 (observed 2026-08-31); "feedly vs inoreader" hits on 3 engines (observed 2026-08-24); "feedly ai" hits on 4 engines, best position 2. Consistent with the direction of this fresh fetch.

3. Competitor search playbook teardown

3.1 Readwise: welding itself into other people's tutorials (the integration ecosystem as distribution)

Readwise has never published any first-hand channel attribution (the honest premise behind the evidence level B of its library case). The observable facts: its positioning is "the middle layer of the highlights pipeline" — importing upstream from Kindle, Instapaper, Pocket and a dozen other sources, syncing downstream to Notion, Obsidian, Roam and Evernote. The founder wrote in 2021 that in the DIY reading pipelines circulating in the PKM community, "Readwise is even a key component" (https://blog.readwise.io/readwise-reading-app/, 2021-09-08, signed by the founder). The effect: anyone writing a Notion/Obsidian workflow tutorial has to mention it — tutorial authors do its SEO for free. Paid acquisition was explicitly rejected: the founder said on HN in 2022 that they had installed Meta/Google pixels for ad experiments but "don't use them anymore" (https://news.ycombinator.com/item?id=34006202, 2022-12-15). The lesson for you: every upstream source or downstream note tool you connect adds a batch of tutorials that will mention you; the integration page is itself a landing page.

3.2 Readwise: absorbing category shutdowns (the named heir in competitors' wills)

Two verifiable events:

• Omnivore (shut down 2024-10-31): the official shutdown announcement explicitly recommended users migrate to "Readwise Reader (paid service)" or Wallabag (open source, self-hosted), and Readwise shipped an Omnivore importer at the same time. Original text in the Web Archive (the original site is 404): http://web.archive.org/web/2024id_/https://blog.omnivore.app/p/details-on-omnivore-shutting-down
• Pocket (shut down 2025): official farewell page at https://getpocket.com/farewell; our library records Readwise founder tristanho showing up in the HN shutdown thread to guide migration paths (checked in library 2026-09-02).

This is not a channel you can book, but the absorption move is copyable: monitor the operating status of free competitors in your category, and within 48 hours of a shutdown announcement ship an importer and show up (on HN and in the announcement comments) to receive the refugees. The fresh data in section 2 proves the demand tail is very long — pocket alternative and omnivore alternative are still in the completions one to two years after the shutdowns.

3.3 Appendix: Inoreader's programmatic comparison pages (library observation, 2026-08-31)

Inoreader's sitemap (3,603 URLs) contains a set of alternative-to-* pages: alternative-to-feedly / flipboard / google-news / apple-news / feedbin / netvibes / newsblur / the-old-reader (8 in total; URLs like inoreader.com/alternative-to-feedly/), plus 357 blog posts and 12 integration pages. This is the cheapest directly copyable play in the category: for every competitor name, alive or dead, build an "alternative to X" landing page — matching, item by item, the autocomplete demand in sections 2.3/2.4.

4. The non-transferable list: what you cannot copy from Readwise

Distilled from the what_not_to_copy field of the readwise case in our library (evidence level B), restated item by item for your "RSS + AI reader" project:

1. The 2017 wedge window is closed. When Readwise started, "no product was doing reading retention" (the founder's own account from 2018), and Kindle highlights were an unoccupied slot. In 2026 you face a category repopulated by survivors after a field of corpses; the cold-start slot for a single-point tool has to be found anew — you cannot assume it is still where it was.
2. Paid-only is not a cold-start strategy. Readwise dared to go paid-only in 2018 because two prior years of freemium had accumulated a die-hard beta user base and evidence of willingness to pay (old beta users even got it free forever at the transition). Charging from day one is not the same thing — the pricing benchmark in section 1 tells you where the endgame price band is, not what to charge in year one.
3. "Default heir after the competitors die off" is an outcome, not a strategy. Omnivore's will named Readwise on the strength of a nine-year survival record at the time. All you can copy is the absorption move in section 3.2 (importer + showing up); the heir status itself cannot be copied.
4. There is no premise that "Readwise makes a lot of money." It has never disclosed any revenue figure (library case: revenue = unknown; the only qualitative fact is the founder's December 2022 statement that a 13-person team runs a sustainable business with revenue 100% from subscriptions). Any decision premised on "the Readwise model is proven lucrative" has no first-hand evidence behind it; the founder's own 2022 footnote reads "Reading retention is a tiny, non-venture scale market. Would not recommend."
5. Nine years of integration ecosystem and archive lock-in cannot be copied as a starting point. Users renew because of years of highlight archives plus the position in the upstream/downstream pipeline (the renewal_reason field in the library); on day one you have neither archive nor pipeline. The play in section 3.1 is a direction, not your current state.

Data completeness statement

• Fresh fetches (2026-09-04): Readwise pricing page re-check, Inoreader pricing page re-check, Feedly pricing page attempt (failed, honestly recorded as unverifiable), Google/Bing autocomplete with 8 seed phrases.
• Library-verified data: all fields of the Readwise case (evidence level B, ten-source research of 2026-09-02), the keyword_profile and search_profile of Inoreader/Feedly (observed 2026-08-24 / 2026-08-31), and the Pocket/Omnivore shutdown facts.
• Explicitly recorded as "not disclosed": all Feedly prices, all Matter prices, the Readwise Lite monthly price, Readwise revenue, and all search-volume figures across the category.

Source URL list

Fresh fetches for this audit (2026-09-04):

1. https://readwise.io/pricing — Lite $5.59 / Full $9.99 (annual) / $12.99 (monthly)
2. https://www.inoreader.com/pricing — Pro $7.50 (annual) / $9.99 (monthly), free tier 150 feeds
3. https://feedly.com/pricing — JS-rendered error template, prices unverifiable
4. https://suggestqueries.google.com/complete/search?client=firefox&q=… — Google autocomplete endpoint (seed phrases: rss reader ai / read it later app / readwise alternative / best rss reader / pocket alternative / omnivore alternative / feedly vs / inoreader vs)
5. https://api.bing.com/osjson.aspx?query=… — Bing autocomplete endpoint (same 8 seed phrases)

Library-verified sources (observation dates on each item):

6. https://readwise.io/read — Reader pricing FAQ (included in Full, 2026-09-02)
7. https://readwise.io/ — homepage "thousands of readers" wording (2026-09-02)
8. https://blog.readwise.io/why-were-bootstrapping-readwise/ — 2018-07-24, the bootstrap + paid-only decision post
9. https://blog.readwise.io/readwise-reading-app/ — 2021-09-08, the "key component" statement, team of 10
10. https://blog.readwise.io/the-next-chapter-of-reader-public-beta/ — 2022-12-14, team of 13, the casualty list, Ghostreader
11. https://news.ycombinator.com/item?id=15339965 — 2017-09-26 Show HN (90 points)
12. https://news.ycombinator.com/item?id=34006202 — 2022-12-15 Show HN (326 points), "100% of our revenue comes from customer subscriptions", ad experiments abandoned
13. https://getpocket.com/farewell — Pocket's official shutdown farewell page (checked 2026-09-02)
14. http://web.archive.org/web/2024id_/https://blog.omnivore.app/p/details-on-omnivore-shutting-down — Omnivore shutdown announcement archive (original site 404)
15. https://www.inoreader.com/alternative-to-feedly/ and 7 other alternative-to pages — library sitemap observation (2026-08-31)
16. Structured data in our library: src/data/readwise.case.ts (evidence level B); the keyword_profile / search_profile of the inoreader and feedly entries in /api/cases (evidence level C)

ProofStack · $99 human evidence audit · 48-hour delivery · at most three per week · your audit redoes all of the data above for your product. Sign up: https://proof-stack-lake.vercel.app/audit

Book your audit ($99, 48-hour delivery, pay after delivery)