# Getting Started

- [Welcome to Katara](/about/getting-started/readme.md): Katara AI helps teams and communities boost productivity with customizable AI workflows, enabling automation, collaboration, and transparent data control.
- [Use Cases](/about/getting-started/use-cases.md): The Katara platform solves use cases by combining AI Agents into pipelines called Workflows. The following use cases can be addressed with Katara Workflows.
- [How does it work?](/about/getting-started/how-does-it-work.md): Katara combines AI agents, workflows, and human oversight to automate tasks with precision, using up-to-date data and ensuring quality through human guidance.
- [Large Language Models](/about/getting-started/how-does-it-work/large-language-models.md): A large language model (LLM) is AI trained to understand and generate human language. Katara AI customizes LLMs like ChatGPT and Claude to power intelligent agents.
- [RAGs - Extending AI context](/about/getting-started/how-does-it-work/rags-extending-ai-context.md): RAGs combine LLMs with external data retrieval to improve accuracy. Katara uses RAGs to power agents with tailored, up-to-date, domain-specific knowledge.
- [Diataxis Framework](/about/getting-started/how-does-it-work/diataxis-framework.md): The Diátaxis framework structures docs into four types—tutorials, how-tos, references, and explanations. Katara uses it to improve content clarity.
- [Agents and Workflows](/about/getting-started/how-does-it-work/agents-and-workflows.md): Agents in Katara combine LLMs with real-time data to handle specific tasks, while workflows link these agents to automate complex, multi-step operations efficiently.
- [People at the helm](/about/getting-started/how-does-it-work/people-at-the-helm.md): Katara AI combines powerful automation with essential human oversight. People ensure quality, handle edge cases, and guide ethical, adaptive improvements across workflows.
- [Data Collection and Usage Policy](/about/getting-started/data-collection-and-usage-policy.md): Katara collects only public, non-sensitive data to support workflows, preserving user anonymity. Data is used solely to analyze trends—not for profiling.