GUIDES · TUTORIALS
Zeus AGI tutorials: practical guides to real work.
Zeus AGI tutorials are short guides to getting real work done with a general intelligence: how to write a goal the loop can hold, how to read a plan, how to check the evidence and how to build on the result. They are written from the product rather than from a script, and every guide assumes you are working on something that matters.
HOW THEY ARE WRITTEN
How these Zeus AGI tutorials are written
Each tutorial starts from a job a person actually has: rebuild this page, check these servers, produce this report. The guide shows the goal as it would be written, the plan you should expect back, the verification that proves the result and the report that closes the loop.
Nothing here is timed, because a general intelligence does not work to a stopwatch and a made-up duration teaches the wrong lesson. Where a guide reaches a step that can fail, it says what failure looks like, so you recognise it in your own work instead of assuming the demonstration was clean. If a guide here cannot say what failure looks like, it is not finished.
THE FIRST GUIDE
A tutorial for your first real task
01
Choose a job with a visible result
Pick something that leaves a trace: a page that should be live, a report that should exist, a test that should pass. A job whose success cannot be inspected cannot teach you how the loop works.
02
Write the goal as a finished state
State the outcome, the constraints and anything that needs your approval. The goal is not a list of steps; it is the description of done, which is what the agent plans against.
03
Read the plan before you approve it
The plan names the tools, the order and the risks. This is the moment to correct a wrong assumption, and it is cheaper than correcting the work later.
04
Inspect the evidence, then keep the skill
Read what was verified rather than only what was claimed. When a job goes well, the procedure can be saved as a skill so the next run starts from experience instead of a blank page. A saved skill is reviewed like code: kept while it earns its place, retired when the work changes.
WHAT THEY COVER
The areas the tutorials cover
The first task
The loop from goal to report, on one job, with the approvals and evidence explained at each step.
Websites and CMS
Building a page, publishing it, working through Agent Bridge and checking the live result.
Code and terminal
Working in a real repository: editing, running tests, reading failures and verifying before a report.
Agent Teams and skills
Splitting a large goal across specialist agents and turning a repeated procedure into a saved skill.
WHAT THEY AVOID
What these tutorials will not do
They will not promise a multiplier on your output, invent a benchmark or show a demo that only works in the recording. They will not hide the cases where the agent asks for approval, stops, or reports a failure, because those moments are where a reader learns what the product actually is.
They also will not pretend that a guide replaces judgement. A tutorial can show the loop closing on one job. Whether that job was the right one to hand over remains your decision.
What a reader should take away is a method rather than a trick: state the finished result, read the plan, inspect the evidence, then decide what to hand over next. That method works on the jobs nobody has written a guide for yet, which is most of them, and it is the main reason these guides stay short.
FAQ
Questions, answered.
Short, direct answers. If yours is missing, the full FAQ page has more.
Short practical guides that show one real job being taken through the loop: goal, plan, execution, verification and report. They are written to be followed while working, not watched, and each one names the points where a run can fail.
No for most guides. The first-task, content, research and media material assumes no code. The code and terminal tutorial does assume you can read a repository listing and run a command, because that is the job being taught.
With the first-task tutorial, then with the guide for the work you actually do. Reading five guides before running one job teaches much less than reading one and testing it on something real, with the evidence in front of you.
The installation and interface guidance is written for any operating system and 11, which is where the desktop application runs. The method, the goal writing, the verification and the Agent Teams guidance apply wherever the product runs.
Yes. Send the job, the systems involved and where you got stuck through the contact page. A workflow that needs a written guide is useful information, and repeated requests are how the tutorial list gets chosen.
BEYOND THIS
This is one example. The next domain is a request away.
Zeus AGI acquires the skills it needs and connects MCP servers, so the scope is not a fixed feature list. If the work is digital or device based, it is in scope.
RELATED
One real job teaches more than five guides.
Pick something with a visible result, write the goal as a finished state and watch the loop close. The capabilities hub shows what else the same intelligence can take on.