loading
loading
Drop your tools (Next.js, Convex, Vercel, MCP, Cursor...) and get a lightweight read on speed, cost, moat, and delivery risk. Useful, not a full audit.
Type tool names one by one (Enter to add each chip) or switch to Paste stack mode and drop your whole setup as free text.
What you get
Try an example
Grade scale
Legend tier
AI coding agent, MCP, evals, and a clean deploy path. Strong setup.
Solid builder
Strong stack. One or two moves from S-tier dominance.
Getting there
Good bones, some tech debt lurking. Improvement path incoming.
Needs a glow-up
LangChain detected. We need to talk.
Scores blend a fixed tool catalog with live model price data — the same feed behind Model Race, cached server-side for up to 12 hours. Every live number shows its as-of date; if the feed is down, the result says heuristic mode plainly. No LLM calls.
Drop tool names as chips or paste your stack. We match common frameworks, AI tooling, deploy infra, and databases.
The result is a lightweight read on Velocity, Cost, Moat, and No-Slop. Use it to decide what to improve next, not as a full engineering audit.
Keep a simple grade card with the verdict and a few upgrade ideas. Then move into the real build tools when you are ready.