Kaer is built around a single agent we call the operator. You describe an outcome in plain language; it plans the work, splits into parallel hands, runs real code, and hands back one reviewable result. Underneath, it routes every piece of work across a stack of frontier models automatically — you never pick a model, and you never manage infrastructure. If you are new, The Operator is the best five-minute read in this guide.
Getting Started
Your first hour with Kaer — from sign-up to your first result
AI & Models
One operator, automatic model routing — no model picking, ever
Search & Research
Live search, source quality, and research synthesis
Agent Workspace
Where the work happens — chat, tasks, and outputs
The platform
The left side of the product gives you the operating map. Dashboard tells you what is active and where attention is needed. Agent is where new work starts — one message in, one outcome out, a shape we call a turn. The work itself happens across surfaces such as Chat, Computer, Files, and Preview, and the operator moves between them as a task needs. As your work expands, Workspaces, Assistants, Automations, Workflows, and Connectors take over.
One operator, many models
Kaer never asks you to choose an AI model. The operator reads each turn and routes it across a stack of frontier models automatically:
- Quick work — lookups, small edits, short drafts — takes fast paths, so simple things stay fast
- Hard work — deep research, complex builds, multi-step reasoning — gets deeper models with more room to think
- Every turn is metered into credits, so you always know what a piece of work cost
You see results, not infrastructure. Learn more in AI & Models.
Learning path
| Stage | Focus | Pages |
|---|---|---|
| Day 1 | One task, one result | Getting Started, The Operator |
| Week 1 | Reusable patterns | Assistants, Workspaces |
| Week 2+ | Automation | Automations, Workflows |
| Ongoing | Scale | Connectors, Team |
What matters early
Start with one-off work in Agent. Once a pattern deserves reuse, capture it in Assistants. When a task should run on its own, move it to Automations or Workflows. Connectors and Mail AI handle delivery outside the dashboard.