VCodeX Docs

Set stream: true to receive the response as server-sent events in the standard OpenAI ChatCompletionChunk format.

resp = client.chat.completions.create(
    model="vcodex/auto",
    messages=[{"role": "user", "content": "explain CRDTs"}],
    stream=True,
)
for chunk in resp:
    print(chunk.choices[0].delta.content or "", end="")

Keepalive

The gateway emits periodic SSE keepalive comments so long-running streams are not cut off by reverse-proxy idle timeouts.

Non-streaming requests return a single JSON ChatCompletion object, exactly as the OpenAI API does.

Streaming