When enabled, the chat will be renamed automatically after the first response from the model. This sends an additional API call to the model and will use some tokens to generate the new name.
Most of the haptic feedback can be disabled by turning this off. This includes feedback from streaming text.
Use this setting to reverse the order of the bottom bar items.
This feature is limited to Geeps Plus users.
Enable this setting to generate or edit images. Currently only OpenAI's DALL-E and GPT-Image models are supported.
When enabled you can generate images or edit existing ones. Image editing is supported with GPT-Image models.
This feature is limited to Geeps Plus users.
When enabled, you can send images to models with vision capabilities (with any provider). You can also attach text based files like PDFs, DOCX (Word), TXT, Markdown and many other formats, including HTML, JSON, XML, CSV, and more.
Image attachments are synced with iCloud across devices. When a chat with image is deleted, the image is also deleted from iCloud.
File attachments are not uploaded to the API, instead they are parsed on device and extracted text is added directly to the chat context. When chats are deleted, text attachments are also deleted.
Choose default vision quality for image attachments. The options are auto, low and high. High quality image processing will use more tokens and cost more.
When using vision or image generation, local image cache is stored on device in the app's Documents directory. Use this option to clear the cache and free up space.
Use these buttons to reset settings or delete the data. Refer to this guide for more information.