DeepFeather

Moats

999 apps · 13 moats

Code stopped being a moat. Every app here is tagged with what's left holding it up — 1 to 3 tags, strongest first.

533of 999 apps (53%) lean on execution polish — the moat AI erodes
67(7%) have nothing else: polish is the entire defence
💅 execution polishPolish, reliability, sync quality, import fidelity — execution, not structure.533 apps →🏗️ infrastructure scaleInfrastructure one person can't match: global hosting, deliverability, uptime, media pipelines.430 apps →🔌 integrationsConnector breadth, and the endless upkeep that keeps every connector working.373 apps →👥 collaborationIt only pays off once the whole team is in it: shared editing, presence, permissions.177 apps →🧠 proprietary modelsCustom-trained or frontier models, plus the compute and inference behind them.132 apps →🎟️ content & rightsLicensed content, media rights, curriculum, template and asset libraries.117 apps →💎 proprietary dataData you can't rebuild: indexes, crawls, live feeds, archives, maps.112 apps →🏛️ compliance & regulationRegulated ground: licensing, payroll, tax, KYC, HIPAA, real legal exposure.107 apps →🕸️ network effectsIt's better because other people are already on it: graphs, communities, audiences.97 apps →🛡️ brand & trustPeople pay because it's this vendor, and nobody got fired for that.83 apps →⛓️ switching costsYour own accumulated history, config and habits make leaving painful.40 apps →🏪 marketplace liquidityTwo sides that need each other, and both of them showed up.17 apps →🔩 hardwarePhysical devices, or data only the vendor's hardware produces.7 apps →

Tags are part of the dataset, so they get argued about in public ·the definitions live in the repo ↗