Published Spaces Chat API

Published Spaces allow organization administrators to expose configured as standalone API endpoints. External applications, custom internal tools, or automated workflows can send chat messages to a published Space and receive AI assistant responses powered by the Space’s instructions, models, and Uploaded Files.

Before invoking a Published Space chat endpoint, a team administrator must publish the Space in Team Settings > Published Spaces. Learn more in Published Spaces.

Endpoint

POST /v1/endpoints/{endpointId}/chat

Request Headers

Header	Type	Required	Description
`X-API-Key`	`string`	Required	Your team API key generated in Developer Tools.
`Content-Type`	`string`	Required	Must be `application/json`.
`omnifact-enable-inline-sources`	`string`	Optional	Set to `"true"` to enable inline source citation formatting in assistant responses.
`omnifact-enable-agentic-workflow`	`string`	Optional	Set to `"true"` to allow agentic multi-tool reasoning during response generation.

Request Body (`application/json`)

{
  "messages": [
    {
      "role": "user",
      "content": "What is our policy on remote work?"
    }
  ],
  "streaming": false
}

Field	Type	Required	Description
`messages`	`array`	Required	Array of message objects (`role`: `"user"` \| `"assistant"`, `content`: `string`).
`streaming`	`boolean`	Optional	Set to `true` for Server-Sent Events (SSE) streaming; `false` for standard JSON (default: `false`).

Response Modes

1. Standard JSON Response (`streaming: false`)

When streaming is false (or omitted), the endpoint processes the entire conversation and returns a single JSON object.

Example Request

curl -X POST "https://connect.omnifact.ai/v1/endpoints/ep_abc123/chat" \
  -H "X-API-Key: your_omnifact_api_key_here" \
  -H "Content-Type: application/json" \
  -H "omnifact-enable-inline-sources: true" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "What are our primary core values?"
      }
    ],
    "streaming": false
  }'

JSON Response (`200 OK`)

{
  "id": "msg_xyz789",
  "role": "assistant",
  "content": "According to our company handbook, our core values are transparency, innovation, and privacy [1].",
  "references": [
    {
      "id": 1,
      "title": "Employee_Handbook_2026.pdf",
      "documentId": "doc_98765"
    }
  ]
}

2. SSE Streaming Response (`streaming: true`)

When streaming is set to true, the server streams the response as real-time Server-Sent Events (SSE) using Content-Type: text/event-stream.

SSE Event Types

The event stream emits typed SSE events as response generation progresses:

Event Name	Description	Payload Structure
`assistant_write`	Streamed text chunk from assistant response.	`{"delta": "text chunk"}`
`references`	Array of source document references cited.	`{"references": [{ "id": 1, "title": "filename.pdf" }]}`
`message_source`	Status indicator for tools/retrieval used.	`{"source": "UPLOADED_FILES", "status": "RETRIEVED", "count": 3}`
`error`	An error occurred during generation.	`{"error": "Description of error"}`
`done`	Stream completed.	`{"done": true}`

Example Streaming Request

curl -N -X POST "https://connect.omnifact.ai/v1/endpoints/ep_abc123/chat" \
  -H "X-API-Key: your_omnifact_api_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Summarize our quarterly goals."
      }
    ],
    "streaming": true
  }'

SSE Stream Output Example

event: message_source
data: {"source":"UPLOADED_FILES","status":"RETRIEVED","count":2}

event: assistant_write
data: {"delta":"Our "}

event: assistant_write
data: {"delta":"quarterly "}

event: assistant_write
data: {"delta":"goals focus on..."}

event: references
data: {"references":[{"id":1,"title":"Q3_Goals.pdf"}]}

event: done
data: {"done":true}

Code Examples

JavaScript / Node.js
Python

const response = await fetch("https://connect.omnifact.ai/v1/endpoints/ep_abc123/chat", {
  method: "POST",
  headers: {
    "X-API-Key": process.env.OMNIFACT_API_KEY,
    "Content-Type": "application/json",
    "omnifact-enable-inline-sources": "true"
  },
  body: JSON.stringify({
    messages: [
      { role: "user", content: "What is our remote work budget allowance?" }
    ],
    streaming: false
  })
});

const data = await response.json();
console.log("Assistant:", data.content);

import requests
import os

url = "https://connect.omnifact.ai/v1/endpoints/ep_abc123/chat"
headers = {
    "X-API-Key": os.getenv("OMNIFACT_API_KEY"),
    "Content-Type": "application/json",
    "omnifact-enable-inline-sources": "true"
}
payload = {
    "messages": [
        {"role": "user", "content": "What is our remote work budget allowance?"}
    ],
    "streaming": False
}

response = requests.post(url, headers=headers, json=payload)
data = response.json()
print("Assistant:", data.get("content"))

Next Steps

Explore document operations with the Documents API
Review Published Spaces documentation to manage endpoint access

Overview

Endpoints

Published Spaces Chat API

Endpoint

Request Headers

Request Body (`application/json`)

Response Modes

1. Standard JSON Response (`streaming: false`)

Example Request

JSON Response (`200 OK`)

2. SSE Streaming Response (`streaming: true`)

SSE Event Types

Example Streaming Request

SSE Stream Output Example

Code Examples

Next Steps

​Endpoint

​Request Headers

​Request Body (application/json)

​Response Modes

​1. Standard JSON Response (streaming: false)

​Example Request

​JSON Response (200 OK)

​2. SSE Streaming Response (streaming: true)

​SSE Event Types

​Example Streaming Request

​SSE Stream Output Example

​Code Examples

​Next Steps

Endpoint

Request Headers

Request Body (`application/json`)

Response Modes

1. Standard JSON Response (`streaming: false`)

Example Request

JSON Response (`200 OK`)

2. SSE Streaming Response (`streaming: true`)

SSE Event Types

Example Streaming Request

SSE Stream Output Example

Code Examples

Next Steps