Xentr Developer Docs

One production order, threaded end to end — create, schedule, run, verify, and trace — using only the REST API.

This guide follows a single order all the way through Factory OS: a run of 1,000 × Bottle Cap 28mm for batch M062026-UBH. Every step is a real endpoint, and the whole flow works with manual quantity logging — no connected machines required. Where the IoT add-on changes a step, a callout shows how.

All requests are scoped to a factory slug and authenticated with a Supabase JWT:

export XENTR_TOKEN="<your-supabase-jwt>"
export FACTORY="penang-plant"   # your factory slug

1. Create the order

An order references a product and a target quantity. The optional batchNumber is your own identifier — it rides along to every downstream record.

curl -X POST https://api.xentr.ai/api/factories/$FACTORY/production-orders \
  -H "Authorization: Bearer $XENTR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "productId": "a1b2c3d4-...",
    "quantityTarget": 1000,
    "batchNumber": "M062026-UBH",
    "plannedStart": "2026-04-20T08:00:00Z",
    "plannedEnd": "2026-04-20T16:00:00Z"
  }'

The order is created in draft. Keep the returned id — every step below uses it as $ORDER.

2. Schedule it

Scheduling assigns the order to a machine (or work area) and moves it to scheduled. Use mode: "single" for one machine; methodId selects the routing.

curl -X POST https://api.xentr.ai/api/factories/$FACTORY/production-orders/$ORDER/schedule \
  -H "Authorization: Bearer $XENTR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "mode": "single", "machineId": "PRESS-01", "methodId": "..." }'

3. Start production

State changes go through one endpoint — /transition with a toState — never by editing a status field. This keeps the audit trail, single-active-order rule, and any quality holds intact.

curl -X POST https://api.xentr.ai/api/factories/$FACTORY/production-orders/$ORDER/transition \
  -H "Authorization: Bearer $XENTR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "toState": "in_production" }'

4. Log output as it runs

On the lite plan you append production counts by hand as parts come off the press. Each entry adds to the running total.

curl -X POST https://api.xentr.ai/api/factories/$FACTORY/production-orders/$ORDER/log-quantity \
  -H "Authorization: Bearer $XENTR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "quantityProduced": 250, "quantityGood": 248, "quantityRejected": 2 }'

With the IoT add-on, you skip this step. When a machine is connected — e.g. a FANUC control via the FOCAS collector — the real part count rolls up automatically against the running order, and log-quantity is only used for manual corrections. The manual path above is the default for the lite plan; the API is identical either way.

5. Complete the run

When the target is met, transition to completed. The order moves into the QA queue.

curl -X POST https://api.xentr.ai/api/factories/$FACTORY/production-orders/$ORDER/transition \
  -H "Authorization: Bearer $XENTR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "toState": "completed" }'

6. Verify quality

A QA reviewer confirms the final good-vs-rejected split and signs the order off. quantityGood + quantityRejected should equal quantityFinal.

curl -X POST https://api.xentr.ai/api/factories/$FACTORY/production-orders/$ORDER/verify \
  -H "Authorization: Bearer $XENTR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "quantityFinal": 1000,
    "quantityGood": 988,
    "quantityRejected": 12,
    "notes": "12 rejects, OD oversize. See NCR-2026-0417-001."
  }'

7. Trace and ship

Every transition the order went through — who moved it, when, and why — is queryable, so the batch carries a complete history from draft to closed.

curl https://api.xentr.ai/api/factories/$FACTORY/production-orders/$ORDER/transitions \
  -H "Authorization: Bearer $XENTR_TOKEN"

That history, the verified quantities, and the batchNumber are the shippable record — no machine telemetry needed to produce it.

Work orders reference

The full order model, splits, and lifecycle.

Quality & andon

Defects, holds, and escalations.

From order to shipped