OpenAI describes the Codex app as a command center for agents. The training turns those capabilities into a practical workflow your team can use every day.
Organize work so Codex receives the right repository, instructions, files, and definition of done without bloated prompts.
Run work in isolated branches or worktrees, inspect diffs, protect uncommitted changes, and keep human review in the loop.
Break a larger project into independent tasks, coordinate agents, avoid overlapping edits, and bring the results back together cleanly.
Turn repeated processes into reusable skills and scheduled automations for triage, reports, checks, documentation, and maintenance.
Set boundaries, choose appropriate approvals, run tests, validate output, and know when a technical expert should review the work.
Use the session to fix a bug, ship a feature, create an internal tool, improve tests, document a system, or standardize your team's Codex playbook.
Product reference: OpenAI's Codex app overview.
Set up Codex around one real project and learn a repeatable build, test, and review loop.
Standardize repository instructions, worktree use, agent delegation, code review, testing, and handoff.
Use Codex for internal tools, data cleanup, reporting, websites, documentation, and technical knowledge work.
Codex is OpenAI's agent for software development and technical work. It can inspect a codebase, change files, run commands and tests, review diffs, and help take a project from an idea to a verified result.
Training is available for developers, product managers, technical founders, agencies, and business operators who want to build internal tools or supervise AI-assisted software work more effectively.
Yes, with the right project and guardrails. Non-developers can use Codex to prototype, document requirements, inspect changes, and build smaller tools. Technical review is still important for production software, security, and sensitive data.
A session can cover the Codex desktop app, local projects, GitHub workflows, skills, worktrees, automations, code review, and how Codex fits alongside ChatGPT Work, Claude Code, Replit, and your existing development stack.
Yes. The most useful sessions start with a real repository, backlog item, internal tool, or repeated technical task. You leave with a working pattern your team can repeat after the session.
Two published sessions, both live on screen share with Evan Weber. Pick the length that matches how much work you want to cover.
$300 USD
Standard offer: 1 to 4 people
Best for one priority workflow. Bring a single task, approved sample inputs, and a clear idea of what a useful result looks like.
$1,000 USD
Standard offer: up to 8 people
Room for several workflows or more practice with each one. The time is shared across the group, so agree on priorities before booking.
Software subscriptions and usage costs are separate from the session fee.
You're welcome to record the session with agreed permissions, and you own that recording. Confirm any trainer-provided recording, follow-up materials or support before booking.
Larger groups, implementation work, or software development are handled as a custom quote: request a team workshop quote or discuss a development project.