1.4
2025-06-12
Updated underlying LLM engines with latest MLX and Llama.cpp source
1.3.1
2025-02-02
Added OpenAI o3-mini support. Improved local DeepSeek R1 support. Added Qwen2.5 7B Instruct 1M to default local models.
Add language support for: Spanish, Chinese, German, Japanese, French, Arabic, Russian, Hindi
1.3
2025-01-27
Smoother text stream.
Tap phone to stop autoscrolling of a message.
Tip Developer Section.
FAQ Section.
Help & Troubleshooting Section.
1.2.1
2025-01-21
Add Deepseek-R1 models. Add settings to allow the app to use more memory. Performance improvements.
1.2
2025-01-02
What’s New in PocketMind
• MLX Support: You can now use both Llama.cpp and MLX engines to run your AI models, giving you more flexibility and power.
• Random Prompt Button: Stuck on what to ask? Tap the “Random Prompt” button to instantly generate creative ideas or choose from a list of 4 default prompts.
• Improved UI: Enjoy a smoother, more intuitive interface that makes navigating and using the app easier than ever.
• Delete Models: Clean up your workspace effortlessly with the ability to delete unused models directly from the app.
Update now to experience these exciting new features!
1.1.2
2024-12-03
Update 1.1.2
- Bug fix
Update 1.1.1
- Added the ability to connect to local servers over http.
This allows the app to connect to local OpenAI compatible servers such as LM Studio.
You will be prompted to allow locally networking the first time you try and connect to a local server.
1.1
2024-11-27
Add support for Online LLM services: OpenAI, Anthropic, xAI Grok and Google Gemini through use of your personal API key