is-agentic-tui

Detect if your code is running inside an agentic TUI (Text User Interface) application like Claude Code, Cursor Agent, Gemini CLI, or Aider.

Installation

npm install is-agentic-tui
# or
pnpm add is-agentic-tui
# or
yarn add is-agentic-tui

Usage

Check if running in any agentic TUI

import { isAgenticTui } from "is-agentic-tui";

if (isAgenticTui()) {
  console.log("Running inside an agentic TUI");
}

Identify which agentic TUI

import { whichAgenticTui } from "is-agentic-tui";

const result = whichAgenticTui();
if (result) {
  console.log(`Running in ${result.tool}`);
  console.log(`Confidence: ${result.confidence}`);
  console.log(`Detected via: ${result.signals.join(", ")}`);
}

Check for a specific tool

import { isSpecificAgenticTui } from "is-agentic-tui";

if (isSpecificAgenticTui("claude-code")) {
  console.log("Running inside Claude Code");
}

Supported Tools

| Tool | Environment Variable | Confidence | |------|---------------------|------------| | Claude Code | CLAUDECODE=1 | high | | Claude Code | CLAUDE_CODE_ENTRYPOINT | medium | | Claude Code | CLAUDE_PATH contains "claude-code" | medium | | Cursor Agent | CURSOR_AGENT=1 | high | | Cursor Agent | CURSOR_INVOKED_AS=cursor-agent | medium | | Gemini CLI | GEMINI_CLI=1 | high | | Aider | AIDER=1 | high* |

*Aider detection is currently unverified and may need adjustment based on actual Aider behavior.

API

isAgenticTui(): boolean

Returns true if running inside any detected agentic TUI.

whichAgenticTui(): DetectionResult | null

Returns detailed detection information, or null if not in an agentic TUI.

import type { DetectionResult } from "is-agentic-tui";
// DetectionResult has: tool, confidence, and signals properties

isSpecificAgenticTui(tool: AgenticTui): boolean

Returns true if running inside the specified tool.

Confidence Levels

• High: Definitive signal detected (e.g., CLAUDECODE=1)
• Medium: Probable signal detected, but less certain (e.g., CLAUDE_CODE_ENTRYPOINT without CLAUDECODE=1)

When multiple tools could match, high-confidence matches take priority.

Use Cases

• Adjust output formatting: Simplify or enrich output for AI consumption
• Skip interactive prompts: Auto-accept defaults when an agent is driving
• Enable verbose logging: Help agents understand what's happening
• Disable animations: Remove spinners and progress bars that don't render well
• Modify error messages: Provide more structured errors for agent parsing

License

ISC