SCOREAI infra
Jev: Independent Index for API Pricing
Jev is an independent index that determines the value of an API answer by calculating expected value across a distribution of probabilities, rather than rounding to a fixed tier, enabling fairer and simpler pricing.
TakeWhat I haven't seen is the buyer being handed the distribution, the confidence and the model version, and told to check the arithmetic. http://demo.tollstile.com/jev
APIs charge for length — tokens, calls, rows. It's a proxy, and it quietly pays a model to pad.
I wanted to see what happens if the price is decided after the work instead. Tollstile already authorizes a ceiling and settles less, so the missing piece was something that could say what an answer was worth, fast enough to sit in the request path. Jev does: typed questions, calibrated probabilities, no prose.
🎲 The part I didn't expect to matter: don't round the answer to a tier. Charging the expected value across the distribution — 0.00 / 0.16 / 0.84 over $0.01 / $0.025 / $0.05 → $0.0179 — is both fairer and simpler than picking a bucket, and sub-cent settlement is already there.
⚡ Ten questions judged in parallel in 598 ms, itemised separately, paid once against one $0.50 ceiling. Output tokens are free, so three typed questions per judgement cost the same as one.
🧾 Outcome pricing isn't new — support tools bill per resolution. What I haven't seen is the buyer being handed the distribution, the confidence and the model version, and told to check the arithmetic.
demo.tollstile.com/jev