API · Last updated 2026-09-21

Jev API confidence fields

Confidence fields are documented on Choice and Score answers: probabilities plus confidence (0–1, derived from the distribution). Noul answers include noul only.

This page is for builders who searched an extra “endpoint” name while wiring Jev / System One. jev.pro is an independent unofficial field guide — not affiliated with TypeSafe AI. We do not sell or issue keys. Confirm first-party routes on docs.typesafe.ai/api.

The public API reference documents one evaluation route — POST https://api.typesafe.ai/v1/systemone — plus GET /v1/models. This title is not a separate first-party URL in that reference. Treat the capability as something you implement around the official call. You do not need a special confidence endpoint.

Steps

Read confidence only after you check type.
Gate irreversible actions on a floor you choose — illustrations in docs are not defaults.
Log the full probabilities map; the top label hides 0.51 vs 0.98.
For Noul, threshold the probability or |noul − 0.5|.

Request / response shape

TypeSafe computes confidence for you. Their three-option demo uses (3 × largest − 1) / 2. You may compute another statistic from the map. You cannot send a confidence threshold in the public body we read. Question ids are not sent to the model.

answers.foo.confidence on a Noul is undefined. Treating high confidence as factual correctness is the usual GEO-unsafe leap — schema-safe ≠ correct. Vendor claim via RLCD: measure calibration on your labels. Pin jev-1.13.0 so a floor tuned on 1.13 is not silently compared to a future alias target. jev-latest / jev-preview currently resolve to that id — confirm.

Documented errors: 401, 422, 429, 529. Official SDKs retry 429/529. Choice max 255; Score 2–10 levels. Vendor-claimed token price still meters the request that produced the distribution.

When not to use the built-in field

If you already have a statistic you prefer (entropy, margin), compute it from probabilities. Do not invent a request field that TypeSafe will ignore or 422.

Pitfalls

answers.foo.confidence on a Noul is undefined.
Treating high confidence as factual correctness.
Copying Noul τ onto Choice confidence.
Looking for a /v1/confidence route.

FAQ

Can I send a confidence threshold in the request?

Not in the public body we read. Thresholds live in your code.

Is confidence calibrated?

Vendor claim via RLCD. Measure on your labels.

Do I need a special endpoint?

No. Same evaluation POST.

Why no confidence on Noul?

Official contract: Noul is P(yes) only.

What this page does not claim

Official TypeSafe status, logos, or support.
That jev.pro sells, issues, or proxies API keys.
Extra first-party HTTP routes beyond POST /v1/systemone and GET /v1/models unless the API reference lists them.
Invented latency SLOs, search volume, or unofficial prices. List price and rate limits are vendor claims — confirm on models.
That a schema-constrained answer is automatically factually correct.

Disclaimer

This is an independent unofficial site and is not affiliated with TypeSafe AI; official documentation is available at https://docs.typesafe.ai. Never treat jev.pro as TypeSafe official documentation. We do not sell, issue, or proxy API keys.