Low-Level Agentic Building Blocks

The Vector Stores API helps agent developers build agents by converting documents into searchable embeddings and storing them in a managed index. This enables agents to quickly retrieve the most relevant passages for a user’s question and use that context to generate more accurate, up-to-date, and organization-specific responses. Developers can use this capability to build scalable knowledge-base features, making agents more reliable and easier to maintain.

API

• Vector Stores API
• Vector Store File Batches API
• Vector Store Files API

• Creating a Vector Store
• Listing Vector Stores
• Using a Vector Store
• Try Semantic Search

The Connector API allows agent developers to use data stored in Object Storage buckets with their Vector Stores. The Vector Store Connector provides a lightweight, configurable way to continuously synchronize data from Object Storage into a Vector Store at scale. This ensures that agents have access to richer and more up-to-date knowledge, improving answer quality and reliability.

• Vector Store Connectors
• Sync Data to a Vector Store

The Files API allows agent developers to securely upload and manage documents and other data that agents can use during task execution, making it easier to ground responses in reliable, project-specific information. By attaching files to workflows such as retrieval and summarization, agents can work with large or complex source materials without requiring content to be manually pasted into prompts. This enables more capable agents that can work with knowledge bases, reports, logs, and datasets in a scalable and reusable way, while keeping data handling centralized and consistent.

API

• Files API

The Containers API provides agent developers with secure, stateful sandboxes for isolated code execution and data processing. By offering a managed environment where agents can run Python scripts and manipulate data, it enables developers to build agents that solve complex problems through logic and computation, rather than relying solely on text generation.

API

Container Files allow agents to persist and manage data directly within execution environments. Developers can programmatically upload source datasets, enable agents to generate outputs such as charts or logs, and retrieve those results later. This localized file system keeps data accessible across multiple steps of a task, enabling seamless multi-stage workflows without manual data transfer.

API