Skip to main content
This page takes you from nothing to a working ChatCLI. Every step links to the full guide when you want the details.
1

Install ChatCLI

Check that it runs:
Building from source, Helm and every other option: Installation.
2

Connect a provider

ChatCLI reads its settings from a .env file: ./.env first, then ~/.chatcli/.env, then ~/.env. Pick a provider and add its key:
Have a ChatGPT, Claude or GitHub Copilot plan? Skip the key: start ChatCLI and run /auth login openai-codex, /auth login anthropic or /auth login github-copilot. See OAuth authentication.
Each provider has a default model; set <PROVIDER>_MODEL to choose another. All fourteen providers and their models are in Supported models.
3

Start your first chat

Type a question and press Enter. Inside the prompt, /config shows the active provider and model, /switch changes them without a restart, and /help lists every command.
4

Give it your project's context

@ commands pull your environment into the prompt:
More in Context commands.
5

Hand it a coding task

Coder mode reads, edits, runs tests and rolls back in a loop, asking before anything risky:
See Coder mode for the approval flow and safety rules.

​
Next steps

Basic usage

The prompt, essential commands and keyboard navigation.

Agent mode

Delegate multi-step tasks that run commands with your approval.

Persistent context

Attach a project once and reuse it across sessions.

Cookbook

Step-by-step recipes for real workflows.