🏰 Software moats

976 apps · 13 moats

Generated code weakens execution polish as a defense. Every app here carries one to three dataset tags describing what still holds it up, strongest first.

536of 976 apps (55%) lean on execution polish, the defense generated code weakens

66(7%) have nothing else: polish is the entire defence

💅 execution polishPolish, reliability, sync quality, import fidelity — execution, not structure.536 apps →🏗️ infrastructure scaleInfrastructure one person can't match: global hosting, deliverability, uptime, media pipelines.432 apps →🔌 integrationsConnector breadth, and the endless upkeep that keeps every connector working.379 apps →👥 collaborationIt only pays off once the whole team is in it: shared editing, presence, permissions.171 apps →🧠 proprietary modelsCustom-trained or frontier models, plus the compute and inference behind them.126 apps →💎 proprietary dataData you can't rebuild: indexes, crawls, live feeds, archives, maps.109 apps →🎟️ content & rightsLicensed content, media rights, curriculum, template and asset libraries.106 apps →🏛️ compliance & regulationRegulated ground: licensing, payroll, tax, KYC, HIPAA, real legal exposure.102 apps →🕸️ network effectsIt's better because other people are already on it: graphs, communities, audiences.81 apps →🛡️ brand & trustPeople pay because it's this vendor, and nobody got fired for that.75 apps →⛓️ switching costsYour own accumulated history, config and habits make leaving painful.38 apps →🔩 hardwarePhysical devices, or data only the vendor's hardware produces.7 apps →🏪 marketplace liquidityTwo sides that need each other, and both of them showed up.6 apps →

Tags are part of the open dataset and can be challenged through the source repository ·the definitions live in the repo ↗