OpenAgereTerminal-native AI coding agent

Read code, edit safely, run commands, and automate engineering work from your terminal.

npm i -g openagere

Terminal-native workflow

Work inside a focused TUI with streaming responses, syntax-highlighted diffs, inline approvals, and keyboard-first controls.

Start a session

Codebase-aware changes

Let the agent search files, inspect context, apply structured patches, and iterate on implementation tasks with your review.

Explore workflows

Configurable safety

Choose read-only, workspace-write, or full-access modes, then control risky tools through clear approval rules.

Review security

Headless automation

Use openagere exec for scripts, CI, build-log triage, batch refactors, and repeatable one-shot agent runs.

Use exec mode

Skills and plugins

Package team-specific instructions, tool integrations, and reusable workflows so OpenAgere adapts to your stack.

Extend OpenAgere

MCP-ready integrations

Connect Model Context Protocol servers, expose approved tools, and bring external context into coding sessions.

Connect tools

Agent delegation

Spawn scoped sub-agents for parallel investigation, implementation, or verification while keeping control in one session.

Learn delegation

Multi-model configuration

Configure OpenAI, Anthropic Claude, or OpenAI-compatible providers, then switch models to fit each task.

Configure models

Why OpenAgere

OpenAgere is an open-source coding agent built for developers who want AI assistance without leaving the terminal. It combines a practical TUI, a structured patch workflow, command execution, code search, approvals, and extensibility into one Rust-powered CLI.

What you can do with it

Ask OpenAgere to explain an unfamiliar repository and identify the files that matter.
Refactor modules, update documentation, and keep changes reviewable through patches.
Pipe logs, diffs, or compiler output into openagere exec for automated analysis.
Connect MCP servers and plugins when your workflow needs external tools or data.
Keep agency and safety explicit with access modes, sandboxing, and approval policies.