Skip to content

Tutorials

Tutorials walk you through Takopi step-by-step. Follow them in order if you're new.

If you already know what you want ("enable topics", "use worktrees"), jump to How-to.

Prerequisites

Before starting, make sure you have:

  • A Telegram account
  • Python 3.14+ and uv (install uv)
  • At least one agent CLI on your PATH:
Agent Install
Codex npm install -g @openai/codex
Claude Code npm install -g @anthropic-ai/claude-code
OpenCode npm install -g opencode-ai@latest
Pi npm install -g @mariozechner/pi-coding-agent

You only need one to get started. Takopi auto-detects what's available.

The tutorials

1. Install and onboard

Set up Takopi, create a Telegram bot, and generate your config.

Time: ~5 minutes

Start here →

2. First run

Send your first task, watch it stream, and learn the core loop: run → continue → cancel.

Time: ~10 minutes

Continue →

3. Projects and branches

Register a repo as a project so you can target it from anywhere. Run tasks on feature branches without leaving your main worktree.

Time: ~10 minutes

Continue →

4. Multi-engine workflows

Use different agents for different tasks. Set defaults per chat or topic.

Time: ~5 minutes

Continue →

What you'll build

By the end of these tutorials, you'll have:

~/.takopi/takopi.toml
├── bot_token + chat_id configured
├── default_engine set
└── projects.your-repo registered

And you'll know how to:

  • Send tasks from Telegram and watch progress stream
  • Continue conversations by replying
  • Cancel runs mid-flight
  • Target specific repos and branches
  • Switch between agents on the fly

After the tutorials

  • How-to guides — goal-oriented recipes (topics, file transfer, voice notes)
  • Reference — exact config keys, commands, and contracts
  • Explanation — architecture and design rationale