You can now quickly turn documents into searchable data using new Azure Logic Apps workflow templates designed to streamline semantic search in Azure Cosmos DB for NoSQL. With just a few clicks, you can extract text from files—including PDFs, Microsoft Office documents, images, markdown, and plain text—then chunk that content, generate vector embeddings using Azure OpenAI, and insert it into your Cosmos DB collection for both vector similarity and full-text search.
These prebuilt workflows help you pull in documents from Azure Blob Storage or Microsoft SharePoint, then use either a standard text parser or Azure Document Intelligence to extract readable text—even from scanned images or PDFs using optical character recognition (OCR). This new capability simplifies a previously complex, multiple-step process and allows you to build GenAI, search, and retrieval augmented generation (RAG) applications faster and more reliably.
Everything runs within your own Azure environment, using your Logic Apps, Azure OpenAI Service, Azure AI Document Intelligence, and Azure Cosmos DB resources, so your data stays in your control.
Templates are now available in the Logic Apps template gallery or directly from the Azure portal for any Azure Cosmos DB for NoSQL account.
Learn more.