Altair

Agent Runtime Manager

Manage, configure, and operate autonomous coding agents with multi-provider support and workspace isolation

Download Now

Key Features

Agent Lifecycle Management

Create, start, stop, and destroy coding agents with real-time status tracking

Multi-Provider Support

Connect to official Anthropic API, third-party endpoints, and custom base URLs

Workspace Isolation

Deterministic workspace boundaries enforced at filesystem level

Monitor Agent

Read-only AI assistant that analyzes chat history and provides insights

Three-Layer Agent Model

Identity Layer (Immutable)

Core agent identification that cannot be changed after creation

Agent ID Agent Type Workspace Path Creation Timestamp

Model Interface Layer (Restart Required)

AI model and provider configuration settings

Provider Selection Base URL & Headers Model Mapping (opus/sonnet/haiku) Capability Declarations

Runtime Parameters Layer (Restart Required)

Behavioral settings that affect agent operation

Auto-edit Settings Planning Mode Git Safety Checks

Getting Started

Install Altair

Download from releases page

Configure Provider

Set API key or environment variable

Create Agent

Specify workspace and type

Start Coding

Launch agent and begin work

UI Layout

Left Column - Agent Management

Middle - Agent Interaction

Right Column - Agent Output

Architecture

Core Runtime (Node.js)

Agent Manager

Model Adapter Layer

UI Shell (Electron/Tauri)

Security

API Key Protection

API keys are never logged and injected only into agent process scope

Workspace Boundaries

Workspace boundaries enforced at filesystem level

Read-Only Monitor

Monitor Agent is strictly read-only with no modification capabilities

Download

Windows (.exe) macOS (.dmg) Linux (.AppImage)

Native Node.js architecture for direct integration with existing tooling