User Guide

Reference documentation for how Cortex works, what its core objects mean, and how to apply the system in practice.

Cortex is an agentic intelligence engine for building systems that continuously collect, filter, and synthesize information into useful outputs. You can use it entirely through the product UI for simpler workflows, or treat it as infrastructure for coding agents and other applications through the API.

The goal of this guide is to make the Cortex object model legible enough that you can design, evaluate, and operate real systems with it, whether you are working inside the UI or integrating Cortex into a broader agent workflow.

Concepts Core Concepts Understand how projects, sources, feeds, digests, and outputs fit together. Reference Glossary Read clear definitions for the core Cortex objects and terms. Examples Examples Browse end-to-end workflow guides and source setup patterns. Reference FAQ Find answers to common setup, troubleshooting, and source-selection questions. Reference Source Types Choose the right connector and learn how supported source types are meant to be used in practice. Reference Digest Types Choose whether a digest should produce a briefing or a dataset and learn the settings that shape each output. Integrations External Credentials Add and manage the provider credentials Cortex uses to connect to external APIs and services. Reference API Keys Create and manage Cortex API keys for scripts, coding agents, and other non-browser callers. API API Reference Open the endpoint-level reference for requests, responses, and schemas.