Streaming

Real-time streaming responses with server-sent events.

How streaming works

OpenCompress supports streaming via server-sent events (SSE), identical to the OpenAI API. Set "stream": true in your request body.

Important: Compression happens before the stream starts. The compression step adds a small fixed latency (~100-300ms), but once streaming begins, tokens arrive at the same speed as a direct API call.

Compression (100-300ms) → Stream starts → Tokens arrive at full speed

Request

curl https://www.opencompress.ai/api/v1/chat/completions \
  -H "Authorization: Bearer sk-occ-your-key-here" \
  -H "Content-Type: application/json" \
  -N \
  -d '{
    "model": "gpt-4o",
    "messages": [{"role": "user", "content": "Tell me a story."}],
    "stream": true
  }'

Response format

Each SSE event contains a JSON chunk:

data: {"id":"gen-abc","object":"chat.completion.chunk","created":1772341560,"model":"gpt-4o","choices":[{"index":0,"delta":{"content":"Once"},"finish_reason":null}]}

data: {"id":"gen-abc","object":"chat.completion.chunk","created":1772341560,"model":"gpt-4o","choices":[{"index":0,"delta":{"content":" upon"},"finish_reason":null}]}

data: {"id":"gen-abc","object":"chat.completion.chunk","created":1772341560,"model":"gpt-4o","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":12,"completion_tokens":85,"total_tokens":97}}

data: [DONE]

The final chunk includes usage data with token counts. Billing is calculated after the stream completes.

SDK examples

Python

from openai import OpenAI

client = OpenAI(
    base_url="https://www.opencompress.ai/api/v1",
    api_key="sk-occ-your-key-here",
)

stream = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Tell me a story."}],
    stream=True,
)

for chunk in stream:
    content = chunk.choices[0].delta.content
    if content:
        print(content, end="", flush=True)