Scored by Claude. Judged by Claude.
The Stork Quadrant
We asked Claude to judge whether Claude could replace every tool in our catalog. Two axes — defensibility and agent-readiness.
Becomes the API
Replaceable as a UI, but kept alive as the API the agents call.
Compounding
Wins twice. Has a real moat AND ships in the agent stack. The strongest position.
Dead Man Walking
An LLM can do most of what this tool's UI promises. No moat, no agent presence.
Sleeping Giant
Has a real moat but invisible to agents. Add an MCP and you'd climb.
Defensibility →
Stork's Picks
Top-scoring tool per quadrant, with Claude's verdict.
Compounding
7Wins twice. Has a real moat AND ships in the agent stack. The strongest position.
Cloudflare is one of the most defensible infrastructure companies on the planet.
MoatsSignalsConfidencelow
- Plaid66
Plaid owns the coordination layer between consumer banking and fintech apps — a role no LLM can replicate alone.
MoatsSignals MongoDB Atlas is infrastructure, not a UI layer — LLMs can write queries against it but can't replace the actual database running in production.
MoatsSignals
Becomes the API
68Replaceable as a UI, but kept alive as the API the agents call.
- Supabase57
Supabase is infrastructure, not a UI — and infrastructure compounds.
MoatsSignalsConfidencehigh
- Sentry54
Sentry's core value is not the analysis — it's the instrumented runtime data that no LLM has access to without Sentry already being in the loop.
MoatsSignalsConfidencehigh
- Algolia53
Algolia's core value is low-latency, globally distributed search infrastructure with a tuning UI — not the intelligence layer.
MoatsSignals Resend's core value is the coordination layer — auth, rate limiting, delivery guarantees, bounce handling, and the ability for agents to send mail reliably at scale without managing SMTP infrastructure.
MoatsSignals
Sleeping Giant
347Has a real moat but invisible to agents. Add an MCP and you'd climb.
AppFolio survives the agent shift because property management is a coordination layer that lives between landlords, tenants, maintenance vendors, and payment systems—no LLM alone can orchestrate that.
MoatsSignalsConfidencehigh
- GitHub59
GitHub is one of the few developer tools that gets stronger as agents proliferate, not weaker.
MoatsSignalsConfidencehigh
Apollo is defensible because autonomous driving requires closed-loop testing on real hardware, regulatory approval from multiple governments, proprietary sensor data and mapping, and orchestration across hardware…
MoatsSignals- Cruise58
Cruise is defensible because autonomous vehicles require physical hardware, regulatory approval (the real moat), continuous operational data from millions of miles, and coordination across insurance, city permits,…
MoatsSignals
Dead Man Walking
5196An LLM can do most of what this tool's UI promises. No moat, no agent presence.
HubSpot's defensibility rests on three real moats: proprietary customer interaction data accumulating across millions of deals, the coordination layer that syncs sales/marketing/support workflows across teams, and…
MoatsSignalsConfidencehigh
- PU
The core value here is not the AI wrapper — it's card issuance, real-time spend controls, and the payment rails underneath.
MoatsSignalsConfidencehigh
Roboflow Annotate's core annotation capability is replaceable — an LLM with vision can generate labels.
MoatsSignalsConfidencehigh
- Twilio43
Twilio is the rails, not the content. An LLM can write the message but cannot send it — carrier relationships, A2P 10DLC registration, short code provisioning, and PSTN interconnects are real regulatory and physical…
MoatsSignals Stripe Billing Trials survives because it owns the payment rails and liability.
MoatsSignals
Scores recompute when Anthropic ships major capabilities.