Configuration¶

All settings use the LAYA_SERVE_ environment prefix (see laya_serve.settings.Settings).

Variable	Default	Meaning
`LAYA_SERVE_BACKEND`	`fake`	`laya` (real weights) or `fake` (weight-free, deterministic)
`LAYA_SERVE_SERVING_MODEL`	`laya-english`	Id reported in the `model` response field
`LAYA_SERVE_EXTRA_MODELS`	``	Extra accepted `model` names, comma-separated
`LAYA_SERVE_DEVICE`	auto	Passed to the Laya `Router` (`cuda`, `cpu`, …)
`LAYA_SERVE_MAX_LOADED`	`1`	Router LRU cap on resident checkpoints
`LAYA_SERVE_PRELOAD`	`false`	Preload all checkpoints (recommended for servers)
`LAYA_SERVE_API_KEY`	unset	When set, requires `Authorization: Bearer <key>`

Backends¶

fake: uniform choice/score distributions, 0.5 for noul. For tests, local API development, and CI. No torch/laya import.
laya: wraps laya.Router (lazy import). Set DEVICE, MAX_LOADED, and PRELOAD=true for servers to avoid cold-start latency.

Unknown LAYA_SERVE_BACKEND values fail fast at startup with a clear error.

Accepted model values out of the box:

jev-latest, jev-preview, jev-1.13.0, jev-1.13, laya, laya-latest, plus the serving model id itself (default laya-english).

Anything else is a 422 with {"error": {"message", "field": "model"}}, so client typos fail fast like Jev. Extend the set without code changes:

LAYA_SERVE_EXTRA_MODELS="jev-1.12,my-alias" laya-serve

The response model field always echoes the serving checkpoint (e.g. laya-english), mirroring how Jev echoes the resolved version id.

Unset LAYA_SERVE_API_KEY means no auth (local dev default). When set, every POST /v1/systemone requires:

curl -H "Authorization: Bearer $LAYA_SERVE_API_KEY" ...

Missing or wrong keys return 401 in the Jev error shape. /healthz and /v1/models stay unauthenticated.