The team have been busy developing the next version of Riku over the past few months and we're excited to today, let users on to test what we have been up with our version 2. Version 2 is a major upgrade in terms of functionality, polish, and user experience. It makes both prompt engineering, fine-tuning, finding, and sharing your AI creations with your team and wider audience even easier than before. We're super excited to introduce V2 of Riku to you all.
Version 2 is all about showcasing the many possibilities of what you can create with AI. We have dashboards for building out prompts, chatbots, prompt chains and fine-tunes with community datasets to get you started on the latter. We're super excited to bring you truly no-code fine-tuning of both OpenAI's and AI21's models.
One of our most popular videos on YouTube involves a showcase of how you can make fine-tunes with OpenAI by using Postman. This got a great reception and many viewers were asking for more content on fine-tuning so we are delighted to now have a truly no-code method.
This new process for fine-tuning could not be more simple. Upload your dataset, give it a name, description and select your technology and adjust the epochs (if you know what epochs are!), and hit the button. It is that simple. The processes will run in the background and your fine-tuned model will be ready to use as soon as OpenAI or AI21 finishes processing it so fine-tuning is a bit part of what is new in V2 but that is not all!
The process of prompt chaining and creating prompt chains has had a major upgrade. It is now easier than ever to choose the prompts to add to a chain and we've introduced community prompt chains to help you learn from the successful creation of others. We've also upped the number of prompts that can be included in a prompt chain to 9.
Riku is a little unique in the way that we provide you the option to create chains of prompts that work in either a sequential or static manner. Sequential prompt chaining is when the output from the first prompt feeds into the second and so on and so on. Static prompt chaining is where every prompt uses the same input data and runs simultaneously, great if you are getting content generated for an ecommerce launch or similar.
Fine-tuning, prompt chaining, prompt engineering, and chatbots are integral parts of V2. Integrations are also super important. From any prompt, you can copy the prompt ID ready to use in any of our integrations with Zapier, Make, Pabbly, Google Sheets, or directly with the Single Endpoint, and from any prompt, you can create a share link that is a fully brandable page for you to share with your audience to generate AI content. How cool is that!
Version 2 is a new look for Riku, a better look, and a more professional finish. We'll be gathering feedback from our early users and fixing any of those initial launch bugs. The team have also been working hard on a new brand identity and website which we'll also have out shortly accompanied by new tutorials, help documentation, and specific landing pages for different types of users.
If you are interested in AI technology and want to get started fine-tuning, prompt engineering and enjoying the best large language models all in a single place consider signing up for Riku today!
