Radar

Radar is a delegation worker for AI coding agents. It lets a high-reasoning host model hand off mechanical work to a cheaper OpenAI-compatible worker model.

Radar is built by ContextRail, the standards layer for AI-ready teams.

The rule is simple:

The host agent thinks. Radar handles the paperwork.

Use Radar for bulk file reading, code search, long-output summarization, and boilerplate drafts. Keep debugging, architecture, security judgment, safety-critical decisions, and exact edits with the host agent.

Start Here

• Getting Started: install Radar, configure a model, and run the first task.
• Delegation Model: understand when to delegate and when to keep work with the host agent.
• MCP Clients: connect Radar to Cursor, Claude Code, Codex, and other MCP clients.
• CLI Reference: look up commands and flags.

Core Workflows

• Documentation updates
• Bulk code reading
• Search before reading
• Boilerplate generation
• Long log summarization

How It Works

Radar packages selected files or content into a stable corpus, sends that corpus to the worker model before the question, and returns a short briefing to the host agent. That preserves the host model's context for judgment and edits.