Documentation

Everything you need to build with Zytherum.

Guides, concepts, and API references for designing, deploying, and governing AI agents. Full documentation is available to customers and design partners.

What the docs cover

Getting started

Set up your workspace and ship a first agent workflow.

Building agents

Design agents, tools, skills, and multi-agent teams.

Workflows & triggers

Chain steps, approvals, and triggers into reliable processes.

Knowledge & memory

Connect documents, databases, and APIs for grounded answers.

Integrations

Connect your systems with connectors and the open protocol.

Evaluation & tracing

Score quality, replay runs, and catch regressions.

Governance & security

Guardrails, approvals, roles, and audit trails.

API reference

Trigger agents and manage resources programmatically.

See it in action

Real runs, recorded from the product.

Every clip below is a real execution — no mockups. Watch how agents, teams, and workflows actually behave, start to finish.

From plain English to a working team

Describe the team you need and the AI composer assembles it on the canvas — a Planner, Sales Analyst, Market Researcher, and Reviewer. Save it, then ask a question in chat and get a real answer back: an insight headline, a data table, a chart, and a recommendation with a confidence level.

Built for developers

Trigger an agent with a single call.

Launch agents and workflows from your own apps over a simple API — with the same governance, tracing, and approvals as the rest of the platform.

Get API access
# run an agent
POST /api/v1/agents/support/run
{
"input": "Customer wants a refund on order #1284",
"mode": "approval"
}
# → 200 · resolved in 3.8s · fully traced