> ## Documentation Index
> Fetch the complete documentation index at: https://docs.eachlabs.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Streaming Overview

> Real-time Server-Sent Events (SSE) streaming for each::sense responses.

## Overview

each::sense uses Server-Sent Events (SSE) for real-time streaming. This enables progressive AI reasoning display, live generation updates, and workflow execution monitoring.

## SSE Format

Each event follows the SSE specification:

```
data: {"type": "event_type", "field": "value"}\n\n
```

The stream terminates with:

```
data: [DONE]\n\n
```

## Enabling Streaming

Set `stream: true` in your request (this is the default):

```bash theme={"dark"}
curl -N -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: YOUR_API_KEY" \
  -d '{
    "messages": [{"role": "user", "content": "Generate a portrait"}],
    "stream": true
  }'
```

## Event Categories

| Category        | Event Types                                                                  | Purpose                    |
| --------------- | ---------------------------------------------------------------------------- | -------------------------- |
| AI Reasoning    | `thinking_delta`, `text_response`                                            | Display AI thought process |
| Tool Operations | `status`, `tool_call`, `message`, `progress`                                 | Show operation progress    |
| Generation      | `generation_response`                                                        | Deliver generated media    |
| Interaction     | `clarification_needed`                                                       | Request user input         |
| Web Search      | `web_search_query`, `web_search_citations`                                   | Search status and results  |
| Workflow        | `workflow_created`, `workflow_fetched`, `workflow_built`, `workflow_updated` | Workflow lifecycle         |
| Execution       | `execution_started`, `execution_progress`, `execution_completed`             | Workflow execution         |
| Terminal        | `complete`, `error`                                                          | Task completion or failure |

## Event Flow: Simple Generation

```
thinking_delta     → "Analyzing request..."
status             → "Searching for best model..."
status             → "Executing nano-banana-pro..."
text_response      → "Here's your portrait!"
generation_response → url: "https://..."
complete           → status: "ok"
[DONE]
```

## OpenAI-Compatible Format

When using `/v1/chat/completions`, events are wrapped in OpenAI-compatible chunks with each::labs extensions in the `eachlabs` field:

```json theme={"dark"}
{
  "id": "chatcmpl-abc123",
  "object": "chat.completion.chunk",
  "model": "eachsense/beta",
  "choices": [{
    "index": 0,
    "delta": {"content": "Generating your image..."},
    "finish_reason": null
  }],
  "eachlabs": {
    "type": "text_response",
    "content": "Generating your image..."
  }
}
```

## Timeouts

| Operation            | Timeout         |
| -------------------- | --------------- |
| HTTP request         | 300 seconds     |
| Streaming connection | 15 minutes idle |
| Image generation     | 10–60 seconds   |
| Video generation     | 60–600 seconds  |
| Workflow execution   | 15 minutes      |