How to Use Vision

Welcome to the Vision feature for Team-GPT, where we’re breaking down the walls between text and images in collaboration. This guide is here to help you understand and use Vision effectively, without getting lost in jargon.

What Is Vision?

Vision is a feature that lets you add images to your conversations on Team-GPT. It uses the latest advancements from GPT-4 with Vision by OpenAI and Claude 3 to understand and analyze these images, enhancing your chats with a rich, multimodal experience.

Features at a Glance

  • Combine Text and Images: Add depth to your conversations by using both words and pictures.
  • Advanced Image Analysis: The platform understands your images, thanks to AI.
  • Easy to Use: Vision is built right into Team-GPT, making it a smooth part of your chat experience.

Getting Started with Vision

What You Need

  • Plan Requirement: Vision is available for Max plan subscribers.
  • Model Compatibility: Make sure you’re using a model that supports Vision (GPT-4V or Claude 3).

How to Use It

  1. Pick a Compatible Model: Before anything else, switch to a Vision-supporting model.
  2. Upload an Image: Click the upload button in the chat. You can add one image per message, up to a 5MB limit.
  3. Ask Your Question: Type any questions or comments related to the image.
  4. Send: The AI will consider both the image and your text to provide a thoughtful response.

Pro Tips

  • Clear and high-quality images get the best results.
  • Text in images should be easy to read without cutting out important visual information.
  • Always switch to a Vision-capable model for image analysis.

Costs: Keeping an Eye on Tokens

Using images in your Team-GPT conversations does use more of your subscription’s resources. Here’s how to keep track:

OpenAI GPT-4 Users

  • Images “cost” based on their size and detail. More detailed views or larger images use more resources.
  • Check out OpenAI’s guide on Vision costs for specifics.

Claude Users

  • Images use tokens, similar to text. The formula tokens = (width px * height px) / 750 helps you estimate the cost if your image is not resized.
  • For price per token and more details, see Claude’s image costs.

General Advice

  • Estimating costs before using Vision can help you manage your resources.
  • Being mindful of image size and detail can help control costs.

Vision brings a fresh dimension to Team-GPT, letting images and text merge for more engaging conversations. Whether you’re breaking down complex ideas or just sharing visuals, it’s a game-changer for collaboration.

Got questions or need a hand getting started? Our support team is here for you. Enjoy the view with Vision!

Updated on March 8, 2024