API · Last updated 2026-09-21

Jev API moderate endpoint

Moderation with Jev is the guardrails cookbook: screen inbound/outbound LLM text with hazard Nouls and severity Scores. You block. There is no documented /v1/moderate.

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.

Steps

1. Put the untrusted text in state.
2. Ask atomic hazard questions (not one “is this ok?” blob).
3. Fail closed on low confidence for irreversible publishes.
4. Keep allowlists deterministic (URLs, user ids).

Map onto the official POST

Same POST. Mix Noul + Score in one questions map. No needs_moderation field is documented on the response. Noul returns noul only; Score returns score, legend, probabilities, confidence. Choice (max 255) can label a hazard class. Question ids are not sent to the model.

Pin jev-1.13.0. Aliases currently resolve there — confirm. Official jaggedness notes: do not believe jev-1.13 is hardened against adversarial state by default. Schema-safe ≠ safe to execute bash. Vendor-claimed context (64k / 32k) still matters; do not paste a whole session if one message is the hazard.

Documented errors: 401, 422, 429, 529. Official SDKs retry 429/529. This site will not sell keys for a moderation demo. TypeSafe does not host a separate moderation SaaS on this page — you write the questions.

When not to use Jev as the only control

Allowlists, authz, and sandboxes stay deterministic. Jev is a judgment layer, not a kernel. If you need OpenAI-style /v1/moderations, that is another vendor.

Pitfalls

• Believing jev-1.13 is hardened against adversarial state by default (jaggedness note).
• Using Jev as the only control on a tool that can execute bash.
• Calling /v1/moderate.
• One blob question instead of atomic hazards.

See also

• API hub — evaluation POST
• LLM guardrails — inbound/outbound screens
• Policy-aware decisions — policy in criteria
• Guardrail — term

FAQ

Is /v1/moderate documented?

No.

Does TypeSafe host a moderation SaaS separate from Jev?

This page is about Jev questions you write.

Can I sell keys here for a moderation demo?

No.

Fail open or closed?

Fail closed on irreversible publishes when confidence is low or Noul is mid-band.

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.