Claude Code training
Claude Code, on your own repo.Not a sample project.
Live sessions for the people who build. How to drive Claude Code on real work, wire it to your systems with MCP, and turn what you keep retyping into skills the whole team runs the same way.
We work in your codebase and your stack. What we build during the sessions stays in your repo.
- Live, on your own codebase
- MCP servers, skills, subagents
- Everything we build stays in your repo
- Led by Omri Dan, AI Expert
Where it stalls
Claude Code is installed.It’s still autocomplete.
It gets used for snippets, and the hard part goes back to a human tab.
Nobody wrote down how the codebase works, so every session relearns it.
It isn’t connected to the issue tracker, the database or the logs.
One person has a setup that works, and it lives on their laptop.
Long tasks get abandoned halfway because nobody trusts the run.
The team knows the tool by name, not by what it can be handed.
So you bought something that builds and use it as a search box.
How we work
We don’t hand you a spec.We build it with you.
Four steps, start to running. You’re in the room for all of them.
01
Mapping
We walk through your day together and find the work worth handing to an agent.
02
Build
Off the call, we build your agent: its memory, your tools connected, the skills for your work.
03
Install
We install it on your computer and connect everything. By the end, it does something real.
04
Workshops
We run your real work with it, side by side, until you watch it do 90% of your repetitive work.
What we cover
Your setup, running by the end.
Every session works on your code and your workflows. Nothing here is a toy example, and nothing is left in a slide.
A project brief Claude Code reads first, so a fresh session already knows your codebase and your conventions.
MCP servers connected to the systems you work in: issue tracker, database, logs, internal docs.
Your repeated tasks written as skills and commands the whole team runs the same way.
Longer runs handled properly: how to scope a task, review the plan, and check the work before it lands.
The review habits that keep an agent honest: what it does alone, and where it stops.
Sessions run live over video against your stack. Everything built in them is committed to your repository.
Testimonials
Business owners who put AI to work, methodically.
Our AI use was scattered. Omri built the system Claude now runs on, in every department, and we ship faster at higher quality.
Ships faster, higher qualityEyal Cafri
VP of R&D · Scytale · Israel
Most of our week was admin: chasing documents, sorting invoices, the same filings. Omri built the systems that take that off, and we handle more clients without growing the staff.
More clients, same staffAndreas
Founder · GMAP CPAs · Cyprus
Managing my design business manually meant working late into the night, and it capped how many clients I could take. Omri put Claude into my workflow, and now most of that runs on its own.
Nights backNoy
Founder · Renovar · Lisbon
Fit
For the people who build.
This is for you if
- You or your team ship code, and Claude Code is already installed.
- You want it working on your real repository, not a sample project.
- You want it connected to your own systems with MCP.
- Someone in-house will own and extend the setup after we leave.
Sounds like you?
Book a free audit callNot the right fit if
- Nobody on the team writes or reviews code.
- You want general AI training for a marketing, ops or finance team.
- You want a recorded course to work through on your own.
- You’d rather not give the tool access to your repository.
If it’s your marketing, ops or finance team that needs AI training, our team AI training is the right page. It’s five weekly sessions for a team of 2 to 5.
FAQ
The questions we actually get.
Ready when you are
Get Claude Code doing real work.
Tell us what your team builds and where the agent keeps stalling. We’ll shape the sessions around that, on your own repository, with your own tools connected.
30 minutes · No slides · No obligation