inference

Blocking text generation (walking-skeleton surface). The model path parameter must use the SIE-safe ID (e.g. Qwen__Qwen3-4B-Instruct); HF-style slashes are rejected with 400.

post/v1/generate/{model}

Path parameters

modelstring required

Model id. Runtime route is Axum catch-all /v1/generate/{*model}; clients using this OpenAPI path template should percent-encode slashes in model ids, for example BAAI%2Fbge-m3.

Headers

X-SIE-MACHINE-PROFILEstring nullable

Preferred GPU or machine profile

X-SIE-Poolstring nullable

Explicit pool routing override

X-SIE-SDK-Versionstring nullable

Client SDK version for skew warnings

Request body

frequency_penaltynumber double nullable
{"stackTrail":"components:schemas:GenerateRequest:properties:grammar","oasType":"schema","type":"unknown","description":"Optional grammar object accepted by the gateway grammar validator."}
max_new_tokensinteger required
presence_penaltynumber double nullable
promptstring required
prompt_cache_keystring nullable
routing_keystring nullable
safety_identifierstring nullable

Sensitive PII - parsed and dropped, never logged or forwarded.

stopstring[] nullable

Stop sequences for the native generate surface.

temperaturenumber float nullable
top_pnumber float nullable

Response

Generated text response

attempt_idstring required
finish_reasonstring required
modelstring required
textstring required
tpot_msnumber double nullable
ttft_msnumber double nullable

Changes