Skip to main content
AI components let you call large language models (LLMs) and other AI services directly from an orchestration pipeline. They cover tasks such as prompting a model, transcribing audio, extracting text from documents, and querying vector databases.

​
Use cases

AI components are useful when you want to enrich pipeline data with generative AI, automate document or audio processing, or build retrieval-augmented pipelines without leaving . Use AI components to:
  • Send a prompt to an LLM and store the response as part of a pipeline, for example to summarize, classify, or generate text.
  • Extract text and structured data from scanned documents, images, or PDFs.
  • Transcribe audio files into text for further processing.
  • Split large documents into smaller chunks ready for embedding or retrieval.
  • Query a vector database to retrieve relevant context for a retrieval-augmented generation (RAG) workflow.
  • Fine-tune a model on your own data so later components can call the resulting custom model.