After a period of development, Haye has officially launched a new memory feature. Unlike common solutions on the market, Haye achieves more precise conversation memory collection with lower resource consumption and introduces multi-dimensional modules such as facts, knowledge, and personal preferences. With each conversation, Haye creates a personalized memory system based on your habits and preferences, helping you efficiently access information during communication.

Additionally, this version introduces a "conversation waiting" feature. You no longer have to wait for the AI to respond before asking more questions; you can now input multiple questions in a row, and all questions will automatically enter a waiting queue. Once the previous question is answered, the system will automatically submit the next question in order, allowing the AI to respond continuously and making communication smoother and more efficient.

Also more improvements, and important fixes:

• Fixed unexpected crash issues with "working with apps".
• Added topic generation model configuration in settings.
• Fixed issues with multi-model message grouping, message appending, and message regeneration.
• Streamlined error message handling for user cancellation
• Merged consecutive reasoning events to prevent duplication for gpt-5 series
• Prevented ESC navigation during Chinese IME composition for smoother input.

Features

• 💬 Ask AI
• 👏 AI Summarize
• 🖊️ AI Write
• 🌍 Translate
• 🔍 AI Search
• 🤔 AI Reasoning
• 🏖️ Style
• 🔧 Function Calling
• 📷 Chat With Image
• 🧾 History
• 📏 Rules for AI
• 🔝 Use API key for custom models
• 🤝 Lifetime Edition
• 💻 Work with app
• 📖 Chat with PDF
• 🚄 Chat Insight
• 🦙 Support for Ollama
• 🦉 Multi-model dialogue
• 🧠 Memory

coming soon

• 🔗 Connectors ...