EkenLab

COURSE 001 · AVAILABLE NOW

Build your first useful AI agent in 2 hours 35 minutes.

Build one runnable meeting follow-up agent with the OpenAI Agents SDK for Python. Turn raw notes into verifiable decisions, actions, owners, deadlines, questions, and source quotes—without inventing missing facts.

Start for $1 Watch the 60-second overview

Beginner+ · Basic Python required · No personal API key required

AGENT RUN / MEETING FOLLOW-UP
RAW NOTES

“Alex will follow up next week… date not confirmed.”

STRUCTURED REPORT

Action item captured. Owner and deadline remain unconfirmed.

VERIFIED QUOTE

Source quote attached to the extracted action item.

SAVED OUTPUT

Typed report written through save_report and ready to inspect.

Beginner+
Basic Python required
No personal API key required
No brief or roadmap required

VERIFICATION PRINCIPLE

Missing facts stay missing.

Source evidence stays attached to extracted items so owners, deadlines, dates, and other facts can be verified instead of invented.

Start Course 001 for $1
Read the 60-second transcript

A meeting ends, and Maya is left with raw notes, unclear actions, and a deadline nobody actually stated. Course 001 is a ready-made Beginner-plus practical course she can start immediately, without a brief or roadmap. In two hours and thirty-five minutes, with Basic Python, she builds one runnable agent using the OpenAI Agents SDK. A Pydantic model structures decisions, actions, owners, deadlines, open questions, and source quotes. A typed save-report tool writes the result through the Agent and Runner loop. Then Maya tests fixtures and edge cases. When an owner is missing, it stays missing. She reproduces a failure, fixes it, and transfers the agent to changed input. The final artifact is hers: code, models, tool, tests, fixtures, saved report, and README. No personal API key is required. Build your meeting follow-up agent. Start Course 001.

WHAT YOU BUILD

One runnable agent. Every important piece remains yours.

The finished artifact includes the code, models, tool, fixtures, tests, saved output, and documentation needed to inspect and adapt the implementation.

Runnable OpenAI Agents SDK agent code
Structured Pydantic output models
Typed save_report tool
Agent and Runner loop
Three fixtures and verification tests
Saved report and README

BEFORE YOU START

Basic Python is required.

You should be comfortable with variables, functions, classes, and running a Python file.

FOCUSED SCOPE

A first implementation—not total mastery.

You do not need a learner-owned API key. The course is designed to help you complete one focused agent implementation, not claim mastery of all agent engineering.

155 MINUTES · SIX PARTS

Build, verify, break, fix, transfer.

1

First Agent

15 min

Run a focused first agent and inspect how the result is produced.

2

Structured output

25 min

Model decisions, actions, owners, deadlines, questions, and source quotes with Pydantic.

3

First tool and agent loop

25 min

Add the typed save_report tool and connect it through an Agent and Runner loop.

4

Own anonymized notes without fabrication

30 min

Use your own safe notes while keeping missing owners, dates, and facts explicitly missing.

5

Tests and edge cases

30 min

Use fixtures, schema checks, quote checks, and an intentionally reproduced failure.

6

Working version and transfer

30 min

Finish the runnable version, document it, and transfer the pattern to a fresh task.

VERIFICATION

Missing facts stay missing. Evidence stays attached.

The agent must not invent owners, deadlines, dates, or other facts. Extracted items remain tied to source quotes so you can inspect what supports them.

Run three distinct fixtures
Check schema validity and attached source quotes
Intentionally reproduce and fix a failure
Complete a transfer task on fresh material

PRICING

Start Course 001 for $1.

Choose a plan through the same EkenLab credit model used across the site. No separate Course 001 credit consumption is claimed here.

FAQ

Questions before you start Course 001.

What Python do I need before starting?

You should understand variables, functions, classes, and how to run a Python file. The course is Beginner+, but it is not a zero-code course.

Do I need my own OpenAI API key?

No. A learner-owned API key is not required for this course.

What does the agent produce?

It turns raw meeting notes into structured decisions, action items, owners, deadlines, open questions, and source quotes, then saves a typed report.

What happens when an owner or deadline is missing?

The agent must leave the fact missing rather than invent it. Extracted items remain tied to source quotes for inspection.

What will I own when I finish?

Runnable agent code, Pydantic models, a typed save_report tool, an Agent and Runner loop, fixtures and tests, a saved report, and a README.

Will this make me an expert in agent engineering?

No. This is a focused first implementation designed to teach one useful agent pattern through building, verification, failure repair, and transfer.

Turn your next meeting into a verified follow-up.

Start for $1 Need a broader path?