System Requirements

• macOS 12.0 (Monterey) or later
• Apple Silicon (M1 or later) — Recommended
• ~500 MB of free disk space
• Internet connection (for AI API access)

Step 1: Download

Visit the official Vell website to download the latest version:

Step 2: Install

System Settings > Privacy & Security — Click "Open Anyway" to allow Vell to run

Step 3: Configure API Key

When you launch Vell for the first time and try to submit a task, you will be prompted to configure an AI provider APIKEY。 API Key Setup section below to get started.

Step 4: Grant Permissions

When Vell needs to access files or folders, it will request your permission. You can manage these in:

• System Settings > Privacy & Security > Files and Folders
• Vell will ask before accessing any folder outside its sandbox
• You can approve or deny each request individually

Welcome Screen

When you first launch Vell, you'll see the home screen with a clean interface:

Vell home screen — task input bar with example prompts

• Task Input Bar — The main text area at the center where you describe what you want to accomplish
• Sidebar — Shows task history and navigation to Skills page and Settings
• Example Prompts — Click to see sample tasks you can try

How to Configure

Settings dialog — select a provider, enter API key, and choose a model

How to Get API Keys

What are Skills?

Skills allow Vell to perform specific types of tasks. Each skill provides a set of tools and instructions that Vell's AI uses to complete tasks in a particular domain. Skills are categorized into different types:

Built-in Skills

Vell comes with essential built-in skills that are always enabled:

• Browser Automation — Navigate websites, fill forms, click buttons, extract data
• Ask User Question — Ask for user input or clarification during task execution
• File Permission — Securely manage file access permissions
• Task Management — Complete tasks, report checkpoints, report thoughts
• Skill Creator — Create new custom skills directly through conversation

From the Recommended List

Creating Skills via Chat

The easiest way to create a custom skill is through conversation:

1. Click 「Create skill in chat」 button on the Skills page
   

   Click 'Create skill in chat' on the Skills page

2. Describe the skill you want to create (e.g., 「Create a skill that organizes my Downloads folder」)
   

   Describe what skill you want — e.g. 'organize my Downloads folder'

3. Vell's AI will generate the skill file and save it to the skills directory
   

   Vell's AI generates the skill and saves it automatically

4. The new skill will appear in your Installed Skills list automatically

Skill File Format

Skills are stored as markdown files in the skills directory. Each skill file follows this structure:

# Skill Name

## Description
What this skill does.

## Instructions
Detailed instructions for the AI on how to use this skill.
Include specific steps, guidelines, and examples.

Example: Weather Skill Structure

Here is an example of how a real skill is structured:

// File: skills/weather/SKILL.md

---
name: weather
description: Weather data retrieval using the Open-Meteo API. Use when Claude needs to get current weather, forecasts, historical data, air quality, or other meteorological information for any location worldwide. This skill provides workflows for querying the free Open-Meteo API with no API key required.
---


  Use the Open-Meteo API at https://api.open-meteo.com/
  - For geocoding: https://geocoding-api.open-meteo.com/v1/search
  - For weather: https://api.open-meteo.com/v1/forecast

Enabling / Disabling Skills

On the Skills page, you can toggle any custom skill on or off using the switch on each skill card. Built-in (system) skills are always enabled and cannot be disabled.

• User skills — Skills you installed or created. You can enable/disable them freely.
• System skills — Built-in skills that are always active (Browser, File Permission, etc.)

Creating a Task

Task Execution View

The execution view shows you everything Vell is doing in real time:

• Messages — Vell's reasoning and actions displayed as a chat conversation
• Thinking Indicator — Shows which tool Vell is currently using (e.g., 「Reading files」, 「Running command」)
• Browser Preview — Toggle the built-in browser view to see Vell's browser actions live
• Todo Sidebar — Shows structured steps Vell has planned for your task
• Permission Requests — Vell asks for your approval before accessing files or performing sensitive actions

Follow-up Questions

After a task completes, you can ask follow-up questions or give additional instructions. Vell will continue working in the same context.

You can also attach files to follow-up messages using the paperclip icon.

How It Works

Vell includes a powerful browser automation engine that can navigate websites, fill forms, extract data, and perform complex web workflows — all visible to you in real time.

What You Can Do

• Search the web and scrape information
• Fill out forms and submit data
• Monitor prices and track products
• Aggregate news and content from multiple sites
• Perform multi-step web workflows

Browser Preview Panel

When Vell is performing browser actions, you can toggle the Browser Preview panel (globe icon in the task header) to watch the automation in real time. This panel shows:

• The live browser view controlled by Vell
• Navigation to URLs and page loading
• Form filling and click actions
• Screenshots taken during the task

Example: Browser Tasks

Viewing History

All your past tasks are listed in the sidebar under the "History" section. Each entry shows the task summary and status.

• Click any task to revisit its execution details
• Task statuses: Queued, Running, Completed, Failed, Cancelled, Stopped

Re-running Tasks

You can re-run any completed task by entering a follow-up message in the execution view. Vell will maintain the context from the previous run.