SilentAPI

Streaming

Receive responses in real time as the model generates them

Streaming sends model output as it is generated. You do not need to wait for the full response. Supported text and multimodal models can also stream function calls.

Enabling Streaming

Set stream: true for the OpenAI Chat Completions API or Anthropic Messages API. Use streamGenerateContent for Gemini. SilentAPI returns server-sent events (SSE).

The Completions API and Responses API return a complete JSON response. They do not support streaming on SilentAPI.

OpenAI Streaming Format

When streaming through /v1/chat/completions, the response is a server-sent events (SSE) stream of OpenAI chat completion chunks. The stream ends with data: [DONE].

curl -N https://api.silentapi.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "gpt-6-sol",
    "messages": [{"role":"user","content":"Write a short poem about the ocean"}],
    "stream": true
  }'

Example events. Cost and usage values are examples:

data: {"id":"chatcmpl_abc","object":"chat.completion.chunk","created":1719584000,"model":"gpt-6-sol","choices":[{"index":0,"delta":{"content":"The"},"finish_reason":null}]}

data: {"id":"chatcmpl_abc","object":"chat.completion.chunk","created":1719584000,"model":"gpt-6-sol","choices":[{"index":0,"delta":{"content":" ocean"},"finish_reason":null}]}

data: {"id":"chatcmpl_abc","object":"chat.completion.chunk","created":1719584000,"model":"gpt-6-sol","choices":[{"index":0,"delta":{},"finish_reason":"stop"}]}

data: {"id":"chatcmpl_abc","object":"chat.completion.chunk","created":1719584000,"model":"gpt-6-sol","choices":[],"usage":{"prompt_tokens":8,"completion_tokens":45,"total_tokens":53},"silentapi":{"cost":0.000038,"fallback":null}}

data: [DONE]

Usage and Cost

The final chunk before [DONE] includes these fields:

Using the OpenAI SDK

from openai import OpenAI

client = OpenAI(api_key="YOUR_API_KEY", base_url="https://api.silentapi.com/v1")

stream = client.chat.completions.create(
    model="gpt-6-sol",
    messages=[{"role": "user", "content": "Write a poem"}],
    stream=True
)

for chunk in stream:
    if chunk.choices and chunk.choices[0].delta.content is not None:
        print(chunk.choices[0].delta.content, end="")

Anthropic Streaming Format

When streaming via /v1/messages, the response is an SSE stream with named events matching the Anthropic streaming specification.

curl -N https://api.silentapi.com/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_API_KEY" \
  -d '{
    "model": "claude-sonnet-5.5",
    "max_tokens": 1024,
    "messages": [{"role":"user","content":"Write a haiku about rain"}],
    "stream": true
  }'

The Anthropic Messages API sends these events in order:

1. message_start: Message metadata.
2. content_block_start: Starts a content block.
3. content_block_delta: Adds content to the block.
4. content_block_stop: Ends the content block.
5. message_delta: Stop reason and usage, with silentapi.cost.
6. message_stop: Ends the stream.

event: message_start
data: {"type":"message_start","message":{"id":"msg_abc","type":"message","role":"assistant","content":[],"usage":{"input_tokens":7,"output_tokens":0}}}

event: content_block_start
data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Soft"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" rain falls"}}

event: content_block_stop
data: {"type":"content_block_stop","index":0}

event: message_delta
data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null},"usage":{"output_tokens":12},"silentapi":{"cost":0.000045,"fallback":null}}

event: message_stop
data: {"type":"message_stop"}

Tool Use Streaming

A function call uses these events:

Using the Anthropic SDK

import anthropic

client = anthropic.Anthropic(
    api_key="YOUR_API_KEY",
    base_url="https://api.silentapi.com"
)

with client.messages.stream(
    model="claude-sonnet-5.5",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Write a haiku"}]
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

Gemini Streaming Format

When streaming through :streamGenerateContent, the response is a server-sent events (SSE) stream of Gemini response chunks.

curl -N "https://api.silentapi.com/v1beta/models/gemini-3.8-flash:streamGenerateContent" \
  -H "Content-Type: application/json" \
  -H "x-goog-api-key: YOUR_API_KEY" \
  -d '{
    "contents": [{ "parts": [{ "text": "Describe a sunset" }] }]
  }'

Each chunk contains candidate parts with incremental text:

data: {"candidates":[{"content":{"parts":[{"text":"The sun"}],"role":"model"}}]}

data: {"candidates":[{"content":{"parts":[{"text":" dips below"}],"role":"model"}}]}

data: {"candidates":[{"content":{"parts":[],"role":"model"},"finishReason":"STOP"}],"usageMetadata":{"promptTokenCount":4,"candidatesTokenCount":32,"totalTokenCount":36},"silentapi":{"cost":0.000021,"fallback":null}}

Using the Google GenAI SDK

from google import genai

client = genai.Client(
    api_key="YOUR_API_KEY",
    http_options={"base_url": "https://api.silentapi.com"}
)

response = client.models.generate_content_stream(
    model="gemini-3.8-flash",
    contents="Describe a sunset"
)

for chunk in response:
    if chunk.text:
        print(chunk.text, end="")

Cross-Format Streaming

SilentAPI translates between streaming formats. A request to /v1/messages returns Anthropic SSE events.

Billing During Streaming

Streaming uses the same model prices as requests without streaming. The cost appears in silentapi.cost in the final event after successful completion.

If a stream is interrupted, SilentAPI charges the reported usage. If usage is absent, it estimates input tokens and output tokens from the content received.

A service failure before output or usage releases the hold. An interrupted stream can close before it sends the cost event.