Unified Chat

This endpoint provides a unified interface for interacting with the AI. It supports general conversation as well as intent-based actions like generating LinkedIn posts. The response is streamed using Server-Sent Events (SSE) to provide real-time feedback, including URL processing status, AI reasoning, and generated content.

URL: POST https://server.feedboss.ai/api/v1/ai-conversation/chat/:workspaceId

Authentication

All endpoints require authentication. You can authenticate using a Service API Key.

Service API Key

Add your Service API Key to the x-api-key header.

x-api-key: fb_live_xxxxxxxxxxxxxxxxxxxxxxxx

---

Path Parameters

Parameter	Type	Description
workspaceId	String	The unique identifier of the workspace (e.g., 691b0d48ad2baa659bf58b4b).

Request Headers

Header	Value	Description
Content-Type	application/json	Required.
Accept	text/event-stream	Recommended to indicate handling of SSE streams.

Request Body

{
  "userInput": "Write a LinkedIn post about the future of AI in marketing.",
  "conversationId": "OPTIONAL_CONVERSATION_ID",
  "contentStyleId": "OPTIONAL_STYLE_ID"
}

Field	Type	Required	Description
userInput	String	Yes	The message or prompt from the user.
conversationId	String	No	The ID of an existing conversation to continue context.
contentStyleId	String	No	The ID of a specific content style or persona to apply to the generation.

Response Format

The response is a Server-Sent Events (SSE) stream. Each event has a type (event name) and data (JSON payload).

Event Types

Keep Alive

• Event: keepalive
• Data: {"ping": true}
• Description: Sent every 15 seconds to keep the connection open.

URL Processing (If URLs are found in input)

• Event: url_processing_start, scraping_url, summarizing_content, extraction_completed
• Data: Contains status messages about scraping and analyzing shared links.

AI Reasoning (Thinking Process)

• Event: reasoning_started
• Event: reasoning_delta -> {"text": "..."} (Chunks of thought process)
• Event: reasoning_ended

Content Streaming (Real-time output)

• Event: chat_delta -> {"text": "..."} (Text chunks for general chat responses)
• Event: content_delta -> {"text": "...", "accumulated": "..."} (Text chunks for generated post content)

Final Results

• Event: chat_response (For General Chat) Data:

  {
    "conversationId": "...",
    "response": "Full response text...",
    "suggestedQuestions": ["..."],
    "conversationHistory": [...]
  }

• Event: post_ids (For Post Generation) Data:

  {
    "conversationId": "...",
    "postId": "...",
    "refinementQuestions": ["Make it shorter", "Add emojis"]
  }

Errors

• Event: error
• Data: {"message": "Error description"}

Example Request (cURL)

curl -X POST http://localhost:5000/api/v1/ai-conversation/chat/691b0d48ad2baa659bf58b4b \
  -H "Content-Type: application/json" \
  -H "x-api-key: fb_live_YOUR_API_KEY_HERE" \
  -d '{
    "userInput": "Create a post about remote work leadership"
  }'

Example Response Stream

event: url_processing_start
data: {"message":"Analyzing shared links...","count":0}
event: reasoning_started
data: {}
event: reasoning_delta
data: {"text":"The user wants"}
event: reasoning_delta
data: {"text":" a post about remote work."}
event: reasoning_ended
data: {}
event: content_delta
data: {"text":"Leading","accumulated":"Leading"}
event: content_delta
data: {"text":" remote","accumulated":"Leading remote"}
...
event: post_ids
data: {"conversationId":"65a1...","postId":"65a2...","refinementQuestions":["..."]}