/eyeroll:init

Set up eyeroll for the current project -- pick a backend, install dependencies, and generate codebase context.

Usage

/eyeroll:init

No arguments needed.

What it does

Step 1: Check installation

Checks if eyeroll CLI is installed. If not, asks which backend you want and installs the appropriate package:

Choice	Install command
Gemini	pip install eyeroll[gemini,download]
OpenAI / OpenRouter / Groq / Grok / Cerebras / openai-compat	pip install eyeroll[openai,download]
Ollama	pip install eyeroll[download]

Step 2: Configure backend

Runs eyeroll init interactively:

• Prompts for backend choice (Gemini, OpenAI, Ollama, OpenRouter, Groq, Grok, Cerebras, openai-compat)
• For Gemini/OpenAI: prompts for API key and validates it
• For Ollama: verifies Ollama is running (or starts it)
• Saves configuration to ~/.eyeroll/.env

Step 3: Generate codebase context

Explores the project and writes .eyeroll/context.md:

1. Reads CLAUDE.md, README.md, pyproject.toml or package.json
2. Checks git log --oneline -10 and current branch
3. Lists root and key source directories
4. Skims 2-3 key source files

The resulting file is a short summary (under 80 lines) that eyeroll uses to ground its analysis in real file paths.

What gets created

File	Location	Purpose
.env	~/.eyeroll/.env	Backend preference and API key
context.md	.eyeroll/context.md	Project summary for grounded analysis

When to re-run

Re-run /eyeroll:init when:

• You want to switch backends
• Your API key has changed
• The project structure has changed significantly (major refactor, new directories)

The codebase context file is what makes the difference between vague guesses and precise file references in eyeroll reports.