README updates

This commit is contained in:
Dillon DuPont
2025-05-28 16:20:15 -04:00
parent 4422b7a0fb
commit f549a25827
2 changed files with 16 additions and 253 deletions

View File

@@ -21,42 +21,33 @@
# 🚀 Quick Start with a Computer-Use Agent UI
### macOS (Local VMs + Cloud Containers)
Get started with a Computer-Use Agent UI with a single command:
**Need to automate desktop tasks? Launch the Computer-Use Agent UI with a single command.**
**macOS:**
```bash
# Requires Python 3.11+
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/scripts/playground.sh)"
```
This script will:
- Prompt you to choose between [C/ua Cloud Containers](https://trycua.com) or local macOS VMs
- Install Lume CLI for VM management (if needed)
- Pull the latest macOS CUA image (if needed)
- Set up Python environment and install/update required packages
- Launch the Computer-Use Agent UI
<details>
<summary>What does this script do?</summary>
**System Requirements:**
- Mac with Apple Silicon (M1/M2/M3/M4 series)
- macOS 15 (Sequoia) or newer
- Disk space for VM images (30GB+ recommended)
### Windows/Linux (Cloud Containers Only)
Get started with a Computer-Use Agent UI with a single command:
1. **Asks if you want to use local VMs?**
- **If yes:**
1. Install VM management CLI
2. `lume pull macos-sequoia-cua:latest` - Download macOS image
3. `lume run macos-sequoia-cua:latest` - Start VM
2. `pip install "cua-computer[all]" "cua-agent[all]"` - Install packages
3. `python -m agent.ui.gradio.app` - Launch UI
</details>
**Windows/Linux:**
```bash
# Requires Python 3.11+ and C/ua Cloud API key
pip install "cua-computer[all]" "cua-agent[all]" ; python -m agent.ui.gradio.app
```
This will:
- Install the required Python packages
- Launch the Computer-Use Agent UI
- Connect to cloud containers through [trycua.com](https://www.trycua.com/)
**System Requirements:**
- Python 3.11+
- Internet connection for cloud container access
*How it works: Computer module provides secure desktops (Lume CLI locally, [C/ua Cloud](https://trycua.com) remotely), Agent module handles local/API agents with OpenAI AgentResponse format and [trajectory tracing](https://trycua.com/trajectory-viewer).*
## Supported [Agent Loops](https://github.com/trycua/cua/blob/main/libs/agent/README.md#agent-loops)
- [UITARS-1.5](https://github.com/trycua/cua/blob/main/libs/agent/README.md#agent-loops) - Run locally on Apple Silicon with MLX, or use cloud providers

View File

@@ -1,228 +0,0 @@
# Quick Start Section Variations
Here are 5 different variations for the Quick Start section, focused on users who want to use Computer-Use Agent UI:
## Variation 1: User-Focused with Technical Context
# 🚀 Quick Start
**Launch the Computer-Use Agent UI in 60 seconds.**
## macOS (Local + Cloud)
```bash
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/scripts/playground.sh)"
```
<details>
<summary>What does this script do?</summary>
The playground script automates the complete setup process:
1. **Install Lume CLI**
```bash
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/libs/lume/scripts/install.sh)"
```
2. **Pull the macOS CUA image**
```bash
lume pull macos-sequoia-cua:latest
```
3. **Run the VM**
```bash
lume run macos-sequoia-cua:latest
```
4. **Install Python packages**
```bash
pip install "cua-computer[all]" "cua-agent[all]"
```
5. **Launch the UI**
```bash
python -m agent.ui.gradio.app
```
You can run these steps manually if you prefer more control over the process.
</details>
## Windows/Linux (Cloud)
```bash
pip install "cua-computer[all]" "cua-agent[all]" ; python -m agent.ui.gradio.app
```
*The Agent UI uses the Computer module to provide secure macOS/Linux desktops via Lume CLI (local) or [C/ua Cloud](https://trycua.com) (cloud), and the Agent module for local/API agents with OpenAI AgentResponse format and [tracing](https://trycua.com/trajectory-viewer).*
---
## Variation 2: Clear User Intent
# 🚀 Quick Start
**Want to use Computer-Use Agents? Get the UI running now.**
### macOS Users
```bash
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/scripts/playground.sh)"
```
<details>
<summary>What does this script do?</summary>
1. **Install Lume CLI for VM management**
```bash
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/libs/lume/scripts/install.sh)"
```
2. **Download the pre-configured macOS image**
```bash
lume pull macos-sequoia-cua:latest
```
3. **Start the virtual machine**
```bash
lume run macos-sequoia-cua:latest
```
4. **Install the Python SDK**
```bash
pip install "cua-computer[all]" "cua-agent[all]"
```
5. **Launch the Computer-Use Agent UI**
```bash
python -m agent.ui.gradio.app
```
</details>
### Windows/Linux Users
```bash
pip install "cua-computer[all]" "cua-agent[all]" ; python -m agent.ui.gradio.app
```
*Technical details: The UI leverages the Computer module (secure desktops via Lume CLI or [C/ua Cloud](https://trycua.com)) and Agent module (local/API agents with OpenAI AgentResponse format and [tracing](https://trycua.com/trajectory-viewer)).*
---
## Variation 3: Direct and Simple
# 🚀 Quick Start
**Get the Computer-Use Agent UI running:**
```bash
# macOS (local + cloud options)
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/scripts/playground.sh)"
```
<details>
<summary>What does this script do?</summary>
1. `curl -fsSL .../install.sh | bash` - Install Lume CLI
2. `lume pull macos-sequoia-cua:latest` - Download VM image
3. `lume run macos-sequoia-cua:latest` - Start VM
4. `pip install "cua-computer[all]" "cua-agent[all]"` - Install packages
5. `python -m agent.ui.gradio.app` - Launch UI
</details>
```bash
# Windows/Linux (cloud containers)
pip install "cua-computer[all]" "cua-agent[all]" ; python -m agent.ui.gradio.app
```
*For developers: Uses Computer module (secure desktops via Lume CLI or [C/ua Cloud](https://trycua.com)) + Agent module (local/API agents with OpenAI AgentResponse and [tracing](https://trycua.com/trajectory-viewer)).*
---
## Variation 4: Problem-Solution
# 🚀 Quick Start
**Need to automate desktop tasks? Launch the Computer-Use Agent UI.**
**macOS:**
```bash
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/scripts/playground.sh)"
```
<details>
<summary>What does this script do?</summary>
Behind the scenes, the playground script runs these commands:
1. **Install Lume CLI**
```bash
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/libs/lume/scripts/install.sh)"
```
2. **Pull macOS CUA image**
```bash
lume pull macos-sequoia-cua:latest
```
3. **Run the virtual machine**
```bash
lume run macos-sequoia-cua:latest
```
4. **Install Python dependencies**
```bash
pip install "cua-computer[all]" "cua-agent[all]"
```
5. **Start the Agent UI**
```bash
python -m agent.ui.gradio.app
```
</details>
**Windows/Linux:**
```bash
pip install "cua-computer[all]" "cua-agent[all]" ; python -m agent.ui.gradio.app
```
*Architecture: Computer module provides secure desktops (Lume CLI locally, [C/ua Cloud](https://trycua.com) remotely), Agent module handles local/API agents with OpenAI AgentResponse format and [tracing](https://trycua.com/trajectory-viewer).*
---
## Variation 5: Ultra Simple
# 🚀 Quick Start
**Start using Computer-Use Agents:**
**macOS:**
```bash
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/scripts/playground.sh)"
```
<details>
<summary>What does this script do?</summary>
1. `lume install` - Install VM management CLI
2. `lume pull macos-sequoia-cua:latest` - Download macOS image
3. `lume run macos-sequoia-cua:latest` - Start VM
4. `pip install "cua-computer[all]" "cua-agent[all]"` - Install packages
5. `python -m agent.ui.gradio.app` - Launch UI
</details>
**Windows/Linux:**
```bash
pip install "cua-computer[all]" "cua-agent[all]" ; python -m agent.ui.gradio.app
```
*Uses Computer module (secure desktops via Lume CLI or [C/ua Cloud](https://trycua.com)) + Agent module (local/API agents with OpenAI AgentResponse and [tracing](https://trycua.com/trajectory-viewer)).*
---
## Recommendation
**Variation 1** is the best choice because:
1. **Clear User Intent**: "Launch the Computer-Use Agent UI" - immediately clear what this does
2. **Time Promise**: "in 60 seconds" sets expectations
3. **Technical Context**: Module details are at the bottom as context, not the focus
4. **User-First**: Focuses on what the user wants to accomplish
5. **Complete**: Still includes all necessary technical information for developers
This puts the user's goal first while keeping the technical architecture details available for those who need them.