ChatGPT Connector

ChatGPT Connector

Conversation history and ChatGPT memories used for raw-memory and persona flows.

Connector Metadata

SDK connector id	chatgpt
Works on	Web, React Native, Swift, Mobile sync
Current collection surface	Web raw memories flow, Mobile sync
Information density	Very high
Signal quality	Very high
Best for	Deep preference context, reasoning style, goals, and raw-memory powered personalization.

Activation

To show this connector in the SDK, add chatgpt to the platform allowlist for the SDK surface you are using.

Web

<OnairosButton
  webpageName="YourApp"
  requestData={['preferences', 'personality']}
  allowedPlatforms={['chatgpt']}
  onComplete={(result) => console.log(result)}
/>

React Native

<OnairosButton
  AppName="YourApp"
  allowedPlatforms={['chatgpt']}
  recommendedPlatforms={['chatgpt']}
  onComplete={(result) => console.log(result)}
/>

Swift

let config = OnairosConfig(
    apiKey: "your_api_key",
    platforms: [.chatgpt],
    recommendedPlatforms: [.chatgpt],
    urlScheme: "yourapp",
    appName: "YourApp"
)

Current SDK Surface

What We Collect

• Conversation IDs, titles, and user-authored messages.
• ChatGPT memory payloads from items, memory_entries, and memories arrays.
• Conversation/message metadata needed for training and retrieval.

Current Onairos Handling

• Conversation history is normalized into the Onairos LLM conversation format.
• ChatGPT memory entries are supported alongside conversation history.
• Mobile sync can import ChatGPT exports into the same normalized format.

Notes

• Supports both conversation history and dedicated memory entries.
• The training dataset pipeline deduplicates conversations by conversation ID before processing.

← Back to Data Sources Overview