tarco Agent CLI

A flexible Agent CLI framework built on top of the Agent Kernel (@tarco/agent). Deploy and run agents with ease, featuring built-in Web UI and powerful extensibility.

Quick Start

Installation

npm install @tarco/agent-cli

Basic Usage

# Start interactive Web UI (default)
tarco

# Run with built-in agents
tarco run agent-tars  # Agent TARS
tarco run omni-tars   # Omni-TARS
tarco run mcp-agent   # MCP Agent

# Run with custom agent
tarco run ./my-agent.js

# Start headless API server
tarco serve

# Headless mode with direct input
tarco --headless --input "Analyze current directory structure"

# Pipeline input
echo "Summarize this code" | tarco --headless

Built-in Agents

tarco CLI includes several built-in agents:

• agent-tars - Agent TARS: Advanced task automation and reasoning system
• omni-tars - Omni-TARS: Multi-modal agent with comprehensive capabilities
• mcp-agent - MCP Agent: Model Context Protocol agent for tool integration

# Use built-in agents
tarco run agent-tars
tarco run omni-tars
tarco run mcp-agent

Core Commands

tarco / tarco run

Launches interactive Web UI for real-time conversation and file browsing.

tarco run --port 8888 --open
tarco run agent-tars --port 8888
tarco run ./my-agent.js --port 8888

tarco serve

Starts headless API server for system integration.

tarco serve --port 8888
# API available at: http://localhost:8888/api/v1/

tarco run --headless

Silent mode execution with stdout output, perfect for scripting.

# Text output (default)
tarco run --headless --input "Analyze files" --format text

# JSON output
tarco run --headless --input "Analyze files" --format json

# Include debug logs
tarco run --headless --input "Analyze files" --include-logs

tarco request

Direct LLM requests for debugging and testing.

tarco request --provider openai --model gpt-4 --body '{"messages":[{"role":"user","content":"Hello"}]}'

tarco workspace

Workspace management utilities.

tarco workspace --init     # Initialize workspace
tarco workspace --open     # Open in VSCode
tarco workspace --status   # Show status

Configuration

Config Files

Supports multiple formats with auto-discovery of tarco.config.{ts,yaml,json}:

// tarco.config.ts
import { AgentAppConfig } from '@tarco/interface';

const config: AgentAppConfig = {
  model: {
    provider: 'openai',
    id: 'gpt-4',
    apiKey: process.env.OPENAI_API_KEY,
  },
  workspace: './workspace',
  server: { port: 8888 },
};

export default config;

CLI Options

# Model configuration
tarco --model.provider openai --model.id gpt-4 --model.apiKey sk-xxx

# Server settings
tarco serve --port 3000

# Workspace path
tarco --workspace ./my-workspace

# Debug mode
tarco --debug

Priority Order

1. CLI arguments (highest)
2. Workspace config
3. User config file (--config)
4. Remote config URL
5. Default config (lowest)

Custom Development

Coming soon

Advanced Features

Event System

Built on event-driven architecture for monitoring agent execution:

const eventStream = agent.getEventStream();

// Subscribe to all events
eventStream.subscribe((event) => {
  console.log('Event:', event.type, event);
});

// Subscribe to specific event types
eventStream.subscribeToTypes(['tool_call', 'tool_result'], (event) => {
  console.log('Tool event:', event);
});

Console Interception

Capture and process console output during execution:

import { ConsoleInterceptor } from '@tarco/agent-cli';

const { result, logs } = await ConsoleInterceptor.run(
  async () => {
    return await agent.run('input');
  },
  {
    silent: true, // Suppress output
    capture: true, // Capture logs
  },
);

Tool & MCP Server Filtering

Filter available tools and MCP servers via configuration:

// In config
const config = {
  tool: {
    include: ['file_*', 'web_*'],
    exclude: ['dangerous_*'],
  },
  mcpServer: {
    include: ['filesystem', 'browser'],
    exclude: ['experimental_*'],
  },
};

# Via CLI
tarco --tool.include "file_*,web_*" --tool.exclude "dangerous_*"
tarco --mcpServer.include "filesystem" --mcpServer.exclude "experimental_*"

API Reference

Refer to TypeScript definitions:

• @tarco/agent-interface - Core agent interfaces
• @tarco/interface - Application layer interfaces
• @tarco/agent-cli - CLI framework interfaces

Examples

[Placeholder: Add screenshots of Web UI interface and CLI usage examples]

Contributing

Welcome to submit issues and pull requests!

1. Follow existing code style
2. Add necessary test cases
3. Update relevant documentation

License

Apache-2.0 License