DuckDuckGo announced yesterday (February 10) that its AI chatbot platform Duck.ai has added real-time voice chat functionality, with a focus on extreme privacy protection.

Different from other voice assistants in the market, the core selling point of this feature lies in its "privacy-first" architectural design. Users can have natural conversations with a large language model (LLM) through an encrypted channel, without worrying about their voice data being monitored or reused by the backend.

image.png

To balance smart experience and data security, DuckDuckGo adopted a unique "middleman" model. Although the underlying intelligence for voice chat is supported by OpenAI, DuckDuckGo has established a firewall between users and OpenAI. The official emphasized that both parties are strictly restricted by contracts: DuckDuckGo will anonymize the audio, and OpenAI is only responsible for processing requests, and is strictly prohibited from retaining data. This means the platform will not store users' chat audio, nor will it use the content for training AI models.

In terms of access barriers, Duck.ai maintains an open strategy: users can experience it for free within daily limits without registering an account. For heavy users, DuckDuckGo introduced a monthly subscription service of $10 (approximately 69.3 RMB), which not only significantly increases usage limits, but also includes additional security features such as personal information removal services and identity theft recovery.

Key Points:

  • 🔒 Privacy Commitment: The official clearly stated that no chat audio will be stored, and all session data will be destroyed immediately after the conversation ends.

  • 🛡️ Data Not Used for Training: Users' voice and AI responses will not be used for any algorithm model iteration training.

  • 🆓 Flexible Usage: Supports direct use of voice functions without registration, while also offering paid subscriptions to meet higher frequency needs for commercial or personal use.