Apps discovered
    Ratings
    232
    NIPs tracked
    30

    Instrument · 01 · Nostr interoperability

    Every Nostr app, measured against the spec.

    Every app discovered from public relays, ranked by real usage, rated NIP-by-NIP with signed ratings anyone can publish and anyone can recompute.

    Read-only crawlsSignature-verifiedMedian of medians

    Readout · 02 · The scoreboard

    Top apps by real usage.

    Full matrix
    Flawless1.0
    Incomplete0.6
    Isolated0.3
    Borked0.1
    self-rated only
    claimed in code, unverified
    no data
    superscript = distinct raters

    Coverage row: share of rated apps at each tier, per NIP.

    Top ten rated apps by monthly authors
    AppMAUOverall010205091017192325465765
    Coverage
    10
    10
    9
    8
    8
    7
    7
    7
    7
    7
    7
    7
    Amethyst2 distinct ratersFFFFFFFFFFFFF
    Bearings1 distinct ratersFFF
    Coracle1 distinct ratersFFFIFFFFFIIIF
    Ditto1 distinct ratersFFFIIFFFFFIF
    Flotilla1 distinct ratersFFFFFF
    Nostria1 distinct ratersFIFFIFIFFIIIF
    nostrudel1 distinct ratersFFFFFFIFFIIII
    snort1 distinct ratersFFFIIFIFFIIIF
    Treasures1 distinct ratersFFFFF
    YakiHonne1 distinct ratersIIIIIFIFIIIIF

    Readout · 03 · Why measure

    Interoperability is the whole point of a protocol.

    Nostr's promise is that your identity, your posts and your relationships work in any app. That promise only holds when apps implement the same specs the same way. Nostrometer makes it measurable.

    01

    Fewer one-client surprises for users

    A zap that works in one app and vanishes in another is a protocol failure that users blame on Nostr. The matrix shows exactly where it happens, so it gets fixed instead of tolerated.

    02

    A concrete fix list for developers

    Every cell below Flawless names a NIP, carries the rater’s note, and comes with an upstream issue filed with the app. Sorted by real usage, it is a backlog with impact built in.

    03

    Pressure that is public and fair

    Nobody is graded in private. Ratings are signed public events, medians are reproducible, the tooling is open, and the number moves when the app does.

    Readout · 04 · How it works

    Three read-only crawls, one rule, and a median.

    1. 01

      Discover

      Crawl kind 31990 handler events from public relays. Verify every signature. Merge duplicates; the developer’s own key wins over a same-named listing from anyone else.

    2. 02

      Rank

      Count distinct authors who tagged their events with the app’s client tag over the last 30 days, via NIP-45 COUNT. A failed query is a ?, never a zero.

    3. 03

      Rate

      Reviews check the current NIP text, then test the round trip against a second client: a cell earns Flawless only when it works on both sides. Anyone can publish a rating; the median per cell sets the tier.

    4. 04

      Publish

      Everything here is read straight from relays or from crawl files you can download. The tooling that produced them is open source.

    Read the full methodology

    Readout · 06 · Updates

    What changed.

    All updates
    No updates published yet. Follow Derek on Nostr for the first one.