Import & Export

Overview

Chatty the Lab supports importing and exporting chat conversations in JSON format. This enables sharing conversations between sessions, backing up data, and migrating chats.

Export format

When exporting a chat, the frontend serializes it to JSON with the following structure:

{
  "model_name": "GPT-4",
  "title": "Medical Q&A",
  "context_text": "You are a medical assistant.",
  "messages": [
    { "role": "system", "content": "You are a medical assistant." },
    { "role": "user", "content": "What is hypertension?" },
    { "role": "assistant", "content": "Hypertension is a condition..." }
  ]
}

Field	Type	Description
`model_name`	`string`	The model used in the original chat
`title`	`string`	Display title of the chat
`context_text`	`string?`	System prompt / context (optional)
`messages`	`array`	Ordered list of messages
`messages[].role`	`string`	One of `system`, `user`, `assistant`
`messages[].content`	`string`	Message text

Downloading a chat

The frontend provides a download button that:

Serializes the active chat’s data (model name, title, context, messages) to JSON
Creates a Blob and triggers a browser download
The file is named after the chat title (e.g., Medical Q&A.json)

No backend call is needed for export — all data is already available in the frontend stores.

Importing a chat

Endpoint

POST /api/sessions/{session_id}/chats/import
Authorization: Bearer <token>
Content-Type: application/json

Request body

{
  "model_name": "GPT-4",
  "title": "Imported: Medical Q&A",
  "context_text": "You are a medical assistant.",
  "messages": [
    { "role": "system", "content": "You are a medical assistant." },
    { "role": "user", "content": "What is hypertension?" },
    { "role": "assistant", "content": "Hypertension is a condition..." }
  ]
}

What happens on import

The backend verifies session ownership
Determines the next available position in the session
Creates a new chats row
Inserts all messages in order, preserving role and content
Returns the full ChatWithMessages object

Response

{
  "id": "uuid-of-new-chat",
  "session_id": "uuid-of-session",
  "position": 2,
  "model_name": "GPT-4",
  "title": "Imported: Medical Q&A",
  "context_text": "You are a medical assistant.",
  "created_at": "2026-03-23T10:00:00Z",
  "messages": [
    {
      "id": "uuid",
      "chat_id": "uuid-of-new-chat",
      "position": 0,
      "role": "system",
      "content": "You are a medical assistant.",
      "created_at": "2026-03-23T10:00:00Z"
    }
  ]
}

Frontend API call

The frontend uses the api.chats.import() method:

const imported = await api.chats.import(token, sessionId, {
  model_name: chatData.model_name,
  title: chatData.title,
  context_text: chatData.context_text,
  messages: chatData.messages
});

Use cases

Backup and restore: Export important conversations and re-import them later
Cross-session sharing: Import a chat from one session into another to continue with a different model set
Collaboration: Share JSON files with colleagues who can import them into their own sessions
Reproducibility: Save and share the exact conversation history for research purposes