Build

To start building with Vectara, review the following sections:

• Data ingestion: Upload documents, configure chunking strategies, structure metadata, and prepare the knowledge base that powers your search and agents. Learn about file formats, ingesting table data, document structure, and metadata filtering.

  Start here if you are setting up your first corpus or need to understand how data flows into Vectara.

• Search and retrieval: Review search methods, configure queries, rerankers, generation parameters, and advanced summarization options. This information helps power both direct queries and agent responses.

  Start here if you understand data ingestion and need to configure how queries work.

• Agents: Create agents with instructions, tools, and sessions. Orchestrate data ingestion and search into governed, conversational AI and document workflow automation.

  Start here if you understand data and search, and you are ready to orchestrate them with agents.

• Pipelines: Automate ingestion from a source system. A pipeline pulls records from a source, runs each one through an agent in its own session, and repeats on a trigger you define. Learn about sources, sync modes, verification, and dead letters.

  Start here if you are ingesting from a source system on a recurring basis rather than uploading files one at a time.

Quick links

• Quickstart - See data ingestion and search work together in about five minutes.
• Agents Quickstart - Create your first AI agent with working code examples.
• API Reference - Complete endpoint documentation.
• Tutorials - Guides for some common use cases.
• Integrations - Connect Vectara to existing tools and workflows.