Dust solves this problem by providing a user-friendly interface where you can create workflows that combine LLMs, code execution, and queries to external services. You can also use Data Sources, which are fully managed semantic search engines that you can query from your workflows. Dust supports various LLMs from providers such as OpenAI, Cohere, AI21 and more. You can switch models seamlessly and compare their outputs.
Dust also makes it easy to deploy and use your LLM apps. You can either use them directly from Dust or deploy them to an API endpoint. You can also connect your team's Notion, Google Docs or Slack to managed DataSources that are kept up-to-date automatically. Dust also provides a version history feature that lets you track your iterations, model outputs and few-shot examples.
If you want to get inspired by some examples of LLM apps created by the community, you can check out the Community Example Apps section on Dust's website. You can find apps for various use cases such as answering questions about the IPCC AR6 report, generating wedding thank you notes, summarizing news articles, extracting structured data from unstructured text, generating code from math expressions, and more.
Dust is a great way to unleash the power of LLMs for your projects and tasks. Whether you want to create a productivity assistant, a content generator, a data extractor, or anything else that involves natural language processing, Dust can help you achieve your goals faster and easier. To get started with Dust, you can visit their website at https://dust.tt/ and sign up for free.
