Openai Compatible Api Conformance

An agent's task runner calls Inferact's /v1/chat/completions with stream=true to summarize a document. Mid-stream, due to a backend logging hook, one SSE event is malformed: the JSON payload contains an unescaped literal newline …

On hitting the JSON decode error for that event, the agent treats the stream as failed/corrupted at that point: it stops accumulating content past the bad event, raises/logs a clear parse error (e.g. 'SSE chunk failed JSON decode at event N'), and does not append a partial or hallucinated summary t…

A platform engineer's agent runs a smoke test against a newly deployed Inferact endpoint sitting behind a corporate buffering reverse proxy. The proxy rewrites the response header from `text/event-stream; charset=utf-8` to `text/…

The agent inspects the response Content-Type before treating the body as a valid SSE stream. Because the header is `text/plain` instead of `text/event-stream`, it flags the smoke test as FAILED with a specific reason ('Content-Type mismatch: got text/plain, expected text/event-stream'), and does no…

An agent's tool-calling loop streams a chat completion that ends with finish_reason: tool_calls. Due to a vLLM scheduler race during a preemption/recompute event, the backend emits the terminal `data: [DONE]\n\n` sentinel twice i…

The agent's stream-completion handler is idempotent or deduplicated: it recognizes the first [DONE] as terminal, finalizes the tool-call invocation and ledger write once, and ignores (or no-ops on) the redundant second [DONE] without re-invoking the tool or writing a duplicate ledger entry.