A perfect custom plan for every team

Using an AI model involves some expenses, and the cost can vary between different models. To make it easier for you to try out various models, each of our plans includes a monthly AI budget. This budget is typically sufficient for most business needs we've seen so far. It's important to remember that this budget resets each month – it doesn't roll over. We add a modest markup to the AI vendor's prices. Our goal is to be clear and upfront about our pricing, and we'll soon introduce a price calculator to help you estimate costs more easily.

For any of our plans, you have the option to purchase a non-expiring budget. This additional budget kicks in only after your regular monthly budget is completely utilized, ensuring continuous AI model usage without interruption.

Prompt chaining is a method to enhance the quality of the final outcome in AI applications. It involves using the output from one prompt as the input for another. This approach is particularly useful when you need to refine the end result. For instance, instead of using a single prompt like "Provide an answer to the question, ensuring it follows the conversation flow," you could use two separate prompts. First, "Provide an answer to the question," and then, "Given this answer, rewrite it so it fits seamlessly into the conversation." This two-step process can lead to more coherent and contextually appropriate results.

These AI blocks are designed to improve their output quality through the process of manually labeling results. This labeling helps the AI learn and adapt for better performance.

An LLM, or Large Language Model, is a type of artificial intelligence model designed to understand, generate, and manipulate human language. These models are "large" both in terms of the size of the neural network (often consisting of billions of parameters) and the vast amount of text data they are trained on.

Whenever your storage capacity has been reached, we remove run-data following the FIFO principle. Rest assured, your labeled data remains untouched, facilitating your experiments with various self-learning AI models.

Only authorized Twin personnel with the necessary permissions can access your data, and this access is strictly limited to maintenance purposes. We enforce robust access controls to ensure the security and confidentiality of your data. Your organization retains full control and ownership, encompassing all workflows and data transmitted through our system.