Questions for legal
This page is not legal advice. It is a list in-house counsel often wants engineers to answer before Jev touches regulated workflows.
Independent unofficial site — we are not counsel, not TypeSafe, and we do not bind anyone. Official Legal: docs.typesafe.ai/legal. Replicas and third-party interfaces named on jev.pro are not Jev and not endorsed.
Affiliation
jev.pro is not TypeSafe. We do not bind TypeSafe. Quotes of their docs can go stale.
Ask engineers
What personal data is in state? Which jurisdictions? Any health/employment/credit? Who is the human in the loop? Can a user contest an auto-action? How long are traces kept?
Ask TypeSafe (via your buyer)
DPA, ZDR, training representations, subprocessors, breach notice — current documents, not a blog.
High-risk list (start from no)
Clinical protocols, hiring decisions, credit, identity proofing, automated legal advice, anything that needs a licensed human. Comparison pages in this batch spell out the split. Counsel still decides.
What counsel should ask TypeSafe (via the buyer)
Engineers can list fields. Counsel still has to ask the vendor — on current paper, not a launch blog:
- DPA / processing addendum — who is controller vs processor for
stateyou POST tohttps://api.typesafe.ai/v1/systemone? - Subprocessors — named list, notice period, region.
- Training representations — TypeSafe’s models page has said customer requests/responses are not used for training. That is a vendor statement. Get the current Legal text, not our paraphrase.
- ZDR / retention — what “zero data retention” covers (logs? traces? support copies?) and for which SKU.
- Breach notice — clock and channel.
- High-risk uses — does the vendor restrict clinical, employment, credit, or identity decisions in the current terms?
- Door split — if traffic also goes through OpenRouter, Vercel, or another gateway, that door is a separate processor. Their card is not TypeSafe’s DPA.
jev.pro cannot answer these. We do not see your traffic and we do not issue keys.
What engineers must put in the briefing packet
Counsel will stall if the packet is “we call an AI.” Bring:
- Exact fields in filtered
state(not the warehouse schema). - Jurisdictions of the data subjects.
- Whether health, employment, credit, or children’s data can appear.
- Who the human in the loop is, and whether a user can contest an auto-action.
- Trace retention and who can query traces.
- The action enum, including fail-closed on timeout / 529.
- Model pin (
jev-1.13.0if floors exist) vsjev-latest. - A sentence that schema-safe ≠ factual correctness — Jev staying inside Choice/Score/Noul does not make the judgment true.
Point counsel at privacy for minimization and questions for buyers for procurement. Comparison pages such as vs legal intake spell product splits; they are not opinions of law.
When not to use this page’s advice
Do not paste this list into a client memo as if jev.pro practiced law. Do not treat our high-risk “start from no” list as a statute. Do not claim “cannot hallucinate” as factual correctness in a warranty. Do not evaluate SemIf/OpenJev under TypeSafe’s DPA — those processors are other parties, not Jev and not endorsed.
What this page does not claim
- Not a legal memo.
FAQ
Can we say Jev cannot hallucinate in a contract? Not as factual correctness. Schema constraint ≠ truth.
Open-source Jev? Hosted Jev is not. Clones are other parties.
Will you review our DPA? No. We publish field-guide questions. Your counsel reads TypeSafe Legal.
Is jev.pro a subprocessor?
No. This site is a public guide. We do not process your production state.
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.
Open-cluster pages are independent field-guide notes. Replicas and third-party interfaces mentioned anywhere on jev.pro are not Jev and not endorsed. Hub: Open. Siblings: privacy guide, questions for buyers, vs legal intake. Canonical: https://docs.typesafe.ai.