Quick Start

This guide will help you get started with Kuse Cowork in minutes.

Step 1: Configure an AI Provider

After launching Kuse Cowork, you need to configure at least one AI provider.

1. Click the Settings icon (⚙️) in the top-right corner
2. Select your preferred AI provider:

Anthropic Claude

1. Get your API key from https://console.anthropic.com/
2. Enter your API key
3. Select a model (e.g., claude-sonnet-4-5-20250929)

OpenAI

1. Get your API key from https://platform.openai.com/
2. Enter your API key
3. Select a model (e.g., gpt-4o)

Local (Ollama)

1. Install Ollama from https://ollama.ai/
2. Pull a model: ollama pull llama3.3:latest
3. Select "Ollama (Local)" as provider
4. No API key needed!

3. Click Test Connection to verify
4. Close the settings panel

Step 2: Start a Chat

The simplest way to use Kuse Cowork is through the Chat interface:

1. Type your message and press Enter
2. The AI will respond with helpful information

Try asking:

• "What can you help me with?"
• "Explain how to use Docker"
• "Write a Python script that prints Hello World"

Example Workflows

Code Review

You: Review the code in src/components/Button.tsx and suggest improvements

Agent: I'll analyze the Button component...
[Tool: read_file src/components/Button.tsx]
[Analysis and suggestions appear]

Bug Investigation

You: The tests in tests/api.test.ts are failing. Help me debug.

Agent: Let me investigate the failing tests...
[Tool: read_file tests/api.test.ts]
[Tool: bash npm test -- tests/api.test.ts]
[Debugging steps and fixes]

Code Generation

You: Create a new API endpoint for user registration in src/routes/

Agent: I'll create a user registration endpoint...
[Tool: read_file src/routes/index.ts]
[Tool: write_file src/routes/auth.ts]
[New file created with registration logic]

Understanding the Interface

Chat View

• Message Input: Type your messages here
• Tool Toggle: Enable/disable tool access
• Project Selector: Choose the working directory

Task View

• Task List: All your tasks in the sidebar
• Progress Panel: Real-time execution status
• Plan View: Step-by-step plan with status indicators
• Tool Output: Results from tool executions

Settings

• Provider Selection: Choose AI provider
• Model Selection: Pick specific models
• API Configuration: Enter API keys and endpoints
• Docker Status: Verify container connectivity

Tips for Better Results

Be Specific

Instead of "fix the bug", try "fix the TypeError in handleSubmit function in src/Form.tsx"

Provide Context

Select a project folder so the agent can read your actual code.

Use Tool Mode for Actions

Enable tools when you want the AI to actually read files or run commands.

Break Down Large Tasks

For complex projects, create multiple focused tasks rather than one large one.

Next Steps

• Configuration Guide - Detailed settings reference
• AI Providers - Learn about supported providers
• Agent System - Deep dive into agent capabilities
• Tools Reference - Available tools and usage