API · Last updated 2026-09-21

Jev API batch endpoint

TypeSafe does document packing many questions into one call (speculative fan-out / parallel questions). That is batching judgments about one state, not a multi-row POST. There is no documented /v1/batch.

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 every question you might need in one questions map.
2. For files of rows, hydrate one state per call — bulk.
3. Pace against vendor RPM / token-per-second defaults (can change).
4. Isolate per-row 422s.

Map onto the official POST

One state per official request. Mix Choice (max 255), Score (2–10), and Noul in that map. Question ids are not sent to the model. Official cookbooks report batching beats one-question-per-call on cost and latency on their example — read the cookbook for current figures; we will not pick a number.

Questions are independent. Unrelated state can still hurt accuracy on jev-1.13. Extra questions are a different warning from extra state distractors. Pin jev-1.13.0. jev-latest / jev-preview currently resolve there — confirm.

Agents default to one-call-per-question — install the agent skill. Documented errors: 401, 422, 429, 529. Official SDKs retry 429/529. Vendor-claimed 1200 RPM and 250k tok/s are ceilings, not a /v1/batch SLA. Vendor-claimed 64k (state+all questions) / 32k (state+longest question) still apply when you fan out. List price is a vendor claim ($0.042 / million input, output free). Schema-safe maps are not factually correct in bulk.

When not to wait for /v1/batch

If you have a CSV of tickets, write a worker. Do not block on a TypeSafe multi-row file API this reference does not list.

Pitfalls

• Calling /v1/batch.
• Extra state distractors (different warning from extra questions).
• One HTTP call per question.
• Sharing one 422 across a file of rows.

See also

• API hub — one state per POST
• Speculative fan-out — many questions
• Batch channel — your file worker
• Bulk imports — row hydration

FAQ

Is /v1/batch documented?

No.

Will extra questions change other answers?

Questions are independent. Unrelated state can still hurt accuracy.

Agent habit?

Install the agent skill.

Is fan-out the same as multi-row batch?

No. Fan-out is many questions, one state.

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.