The Lock-in Detector
Rate your AI stack on five factors — data portability, proprietary formats, egress cost, API standardization, and switching effort — and the detector weights each into a single 0–100 lock-in score with a tier from low to severe. The moat is what it does next: instead of a vague "avoid vendor lock-in" slogan, it names the exact mechanism holding you hostage (a proprietary schema, a client-side format, a punishing egress bill) and the concrete swap that neutralizes it. Runs entirely in your browser (0 uploads, works offline).
How to use it
- Drag each of the five sliders to describe your current stack, or load a preset — Portable, Typical, or Captured.
- Watch the lock-in score and tier update live, and read the per-factor points bar to see which axis is dragging you down.
- Open the X-ray panel to see the named mechanism behind each flagged axis and the specific swap that frees you.
- Copy or share your score — a low sovereignty number is the kind of thing engineers argue about.
What this clears up (the fundamentals)
- Lock-in is a sum of mechanisms, not a mood — it comes from concrete things: a schema only one vendor reads, a wire format no other model speaks, an egress fee priced to punish leaving. Score each and the total is honest.
- Some factors invert — high data portability and open, standardized APIs reduce lock-in, so a great score there pulls the total down. More-is-better and more-is-worse axes are weighted in opposite directions.
- The biggest factor is where you start — decoupling is a project; you do the heaviest axis first. Naming it turns "we should really fix our lock-in" into a next action.
- Open interfaces are the exit — MCP for tools, an OpenAI-compatible gateway for the API, open formats for data. Each one converts a wall into a config change.
Where it's used
A top-of-funnel AI-engineering reliability check — the fastest way to feel why "we might be locked in" is a measurable score, not a worry. It's a miatz build-lab concept playable: play it here, then learn to build the scoring rubric and the structural reveal yourself.
FAQ
What exactly is "vendor lock-in" here?
The cost — in money, time, and rework — of leaving your current AI provider or stack. This tool decomposes it into five weighted factors so the abstract fear becomes a concrete, comparable score.
Why does high data portability lower my score?
Because portability is the opposite of lock-in. For the two "invert" factors (data portability and API standardization), a high rating means you're free to leave, so it subtracts from the total. The three other factors add lock-in as they rise.
How is the score calculated?
Each factor is rated 0–4 and carries a weight (they sum to 100). Invert factors are flipped first, then each contributes weight × (rating / 4) points; the sum, rounded, is your 0–100 score. It's a deterministic rubric — no model call, no randomness.
What does "name the mechanism" mean?
Instead of telling you to "avoid lock-in," the reveal points at the specific thing creating it — a proprietary tool schema, a bespoke API surface, an egress-metered store — and gives the exact swap (MCP, a provider-agnostic gateway, an open export format).
Is anything uploaded?
No. Your ratings are scored entirely in your browser — nothing is transmitted, stored, or logged. Turn off your Wi-Fi and it still works.
Limits
A teaching rubric with five fixed factors and hand-set weights — it's a fast, opinionated read on lock-in, not a procurement audit. Your real exit cost depends on data volume, contracts, and how deeply the stack is wired in. The mechanism it names (open interfaces beat proprietary ones) is exactly the real practice.
Related
Part of the Demystify Playgrounds. Explore the rest from the Playgrounds home.
Bookmark this page (Ctrl+D, or ⌘D on Mac) — it works offline the next time you need it.