Getting Started

OpenAgere is a terminal-native AI coding agent. It can inspect your project, edit files through structured patches, run commands, and coordinate work while keeping permissions explicit.

Install

npm i -g openagere
openagere

See Install OpenAgere for Homebrew, direct release, and source-build options.

Start the TUI

Run OpenAgere from the repository you want it to understand:

cd path/to/your/project
openagere

The terminal UI opens with a chat composer. Type a request in natural language and OpenAgere will gather context, propose actions, apply patches when allowed, and summarize what changed.

Try your first prompts

Explain how this repository is structured.
Find the API route that creates users and describe the validation flow.
Add a --verbose flag to the CLI and update the documentation.
Fix the failing test in src/parser.rs and explain the root cause.

Good prompts give OpenAgere a clear objective, expected files or modules when known, and any constraints you care about.

Understand the loop

1. Plan — OpenAgere breaks larger requests into steps when useful.
2. Inspect — It reads files, searches the codebase, and asks for missing context.
3. Act — It edits via structured patches and runs approved commands.
4. Verify — It can run targeted tests or checks when the environment allows.
5. Report — It summarizes changed files, validation results, and next actions.

Common controls

Key or command	Action
Enter	Submit a prompt
Ctrl+C	Interrupt the current turn
Ctrl+D	Exit gracefully
/model	Switch the active model
/resume	Resume a previous session
/compact	Compress conversation context
/init	Generate or update repository instructions
/help	Show available help

Run one-shot tasks

Use openagere exec when you want a non-interactive run for scripts, CI, or quick analysis:

openagere exec "Summarize this repository and identify the main build command"
git diff HEAD~1 | openagere exec "Write a concise commit message for this diff"

Learn more in Headless Exec.

Next steps

• Configure models and defaults in Configuration.
• Review access modes in Security Model.
• Connect external tools with MCP.
• Package reusable workflows with Skills and Plugins.