pllmResponses API
Use PLLM's Responses API surface and understand its tested conformance scope.
Send Responses API requests to http://127.0.0.1:8080/v1/responses with a bearer
token for the local gateway. The endpoint supports non-streaming and SSE
streaming text responses, stored response retrieval and cancellation,
continuation with previous_response_id, and local function-call items.
POST /v1/responses/compact performs bounded local history compaction. It keeps
recent items and marks elided history; it does not ask a provider to generate a
semantic summary.
Conformance result
The gateway passed all 17 conformance tests from
OpenResponses against its 2026-04-24
specification at pinned upstream commit
92c12d96d7b61d6d15e2214daa5e9c6000ab6e1c.
The run covered the gateway contract exercised by those tests. It does not prove
support for every Responses API field, hosted tool, transport, or model modality.
Current model traffic is text plus local function tools. Image and file inputs are not model-understood modalities, and provider-built-in tools are not available. Unsupported fields fail instead of being silently accepted.
Use the OpenAI Python SDK or OpenAI Agents SDK for tested Python client paths.