Estonia · Since February 2025
Estonia's Claude Code experts
We've been using and learning, and teaching Claude Code — Anthropic's terminal coding agent — since the day it launched in February 2025. 300+ developers trained and a growing Estonian Claude Code community.
Join #claude-code-ee on Discord
while(task) { think → act → observe }The Definition
What is a coding agent?
An LLM that uses tools in a loop with persistent context. It thinks, acts, and observes — until the task is done. Claude Code is the canonical example: it lives in your terminal, reads your codebase, runs tests, and ships changes.
CODING AGENT
LLM + Tools + Loop + Context
while(task) { think → act → observe }IDE AUTOCOMPLETE
Suggest a line, wait for tab
prompt → response → doneWhy Claude Code
The tool developers actually love
Claude Code tops developer satisfaction surveys for a reason. Four reasons, actually.
Anthropic shipped Claude Code in February 2025 — 11 months before Cursor's terminal agent. The harness, hooks, and MCP ecosystem are years ahead.
No IDE sandbox. Full bash, full filesystem, full git. Compose with any CLI tool you already use — gh, docker, kubectl, your own scripts.
Hooks, sub-agents, MCP servers, custom slash commands, skills. The agent is genuinely extensible — not a chatbot with a fancy UI.
Claude Opus and Sonnet lead every coding benchmark that matters. Claude Code is built by the team training the models — tight feedback loop, fast iteration.
Track Record
We've been here since day one
Anthropic releases Claude Code. We start using it in our own work and learning the harness from day one.
Hackathons with Datel, Bitweb, Holm, and Aktiva Finance. Hands-on, practical usage — building real things, not slideware.
Plepic launches the Estonian Claude Code community: meetups, webinars, and structured 6-week training. More hackathons (Aktiva vol2, Delfi Meedia, Thorgate). 300+ developers trained, hackathon average rating 8.4/10.
Cursor shipped terminal agents in Jan 2026. GitHub Copilot CLI followed in Feb 2026. The competition is roughly a year behind Claude Code — and we train the teams who didn't wait.
The Shift
Why we bet on Claude Code
while(task) { think → act → observe }
Every loop iteration is informed by the previous “observe” — that is reflection. The iteration itself creates learning. Claude Code was the first tool to make this loop feel native.
“After writing my first Claude Code loop, I couldn't sleep. Every repetitive task, every debugging session, every code review — I could see how each one could be transformed.
Once you build your first loop, you stop thinking about AI as a chatbot. You start seeing it as a collaborator that improves with each iteration. That's the shift we teach.”
Kaido Koort · Founder, Plepic
Ship larger projects with the same team
Claude Code lets small teams take on work that used to need a much larger headcount. 6 weeks of structured training to assemble the harness around your codebase — tools, context, and workflows that scale with you.
View Full ProgramNot ready to commit? Talk to Kaido about a hackathon for your team.