bytebot README

The project's README from github.com/bytebot-ai/bytebot. As one markdown file: README.md.

Bytebot Logo

Bytebot: Open-Source AI Desktop Agent

bytebot-ai%2Fbytebot | Trendshift

An AI that has its own computer to complete tasks for you

Deploy on Railway

Docker License Discord

🌐 Website • 📚 Documentation • 💬 Discord • 𝕏 Twitter

Deutsch | Español | français | 日本語 | 한국어 | Português | Русский | 中文

https://github.com/user-attachments/assets/f271282a-27a3-43f3-9b99-b34007fdd169

https://github.com/user-attachments/assets/72a43cf2-bd87-44c5-a582-e7cbe176f37f

What is a Desktop Agent?

A desktop agent is an AI that has its own computer. Unlike browser-only agents or traditional RPA tools, Bytebot comes with a full virtual desktop where it can:

Think of it as a virtual employee with their own computer who can see the screen, move the mouse, type on the keyboard, and complete tasks just like a human would.

Why Give AI Its Own Computer?

When AI has access to a complete desktop environment, it unlocks capabilities that aren't possible with browser-only agents or API integrations:

Complete Task Autonomy

Give Bytebot a task like "Download all invoices from our vendor portals and organize them into a folder" and it will:

Process Documents

Upload files directly to Bytebot's desktop and it can:

Use Real Applications

Bytebot isn't limited to web interfaces. It can:

Quick Start

Deploy in 2 Minutes

Option 1: Railway (Easiest) Deploy on Railway

Just click and add your AI provider API key.

Option 2: Docker Compose

git clone https://github.com/bytebot-ai/bytebot.git
cd bytebot

# Add your AI provider key (choose one)
echo "ANTHROPIC_API_KEY=sk-ant-..." > docker/.env
# Or: echo "OPENAI_API_KEY=sk-..." > docker/.env
# Or: echo "GEMINI_API_KEY=..." > docker/.env

docker-compose -f docker/docker-compose.yml up -d

# Open http://localhost:9992

Full deployment guide →

How It Works

Bytebot consists of four integrated components:

  1. Virtual Desktop: A complete Ubuntu Linux environment with pre-installed applications
  2. AI Agent: Understands your tasks and controls the desktop to complete them
  3. Task Interface: Web UI where you create tasks and watch Bytebot work
  4. APIs: REST endpoints for programmatic task creation and desktop control

Key Features

Example Tasks

Basic Examples

"Go to Wikipedia and create a summary of quantum computing"
"Research flights from NYC to London and create a comparison document"
"Take screenshots of the top 5 news websites"

Document Processing

"Read the uploaded contracts.pdf and extract all payment terms and deadlines"
"Process these 5 invoice PDFs and create a summary report"
"Download and analyze the latest financial report and answer: What were the key risks mentioned?"

Multi-Application Workflows

"Download last month's bank statements from our three banks and consolidate them"
"Check all our vendor portals for new invoices and create a summary report"
"Log into our CRM, export the customer list, and update records in the ERP system"

Programmatic Control

Create Tasks via API

import requests

# Simple task
response = requests.post('http://localhost:9991/tasks', json={
    'description': 'Download the latest sales report and create a summary'
})

# Task with file upload
files = {'files': open('contracts.pdf', 'rb')}
response = requests.post('http://localhost:9991/tasks',
    data={'description': 'Review these contracts for important dates'},
    files=files
)

Direct Desktop Control

# Take a screenshot
curl -X POST http://localhost:9990/computer-use \
  -H "Content-Type: application/json" \
  -d '{"action": "screenshot"}'

# Click at specific coordinates
curl -X POST http://localhost:9990/computer-use \
  -H "Content-Type: application/json" \
  -d '{"action": "click_mouse", "coordinate": [500, 300]}'

Full API documentation →

Setting Up Your Desktop Agent

1. Deploy Bytebot

Use one of the deployment methods above to get Bytebot running.

2. Configure the Desktop

Use the Desktop tab in the UI to:

3. Start Giving Tasks

Create tasks in natural language and watch Bytebot complete them using the configured desktop.

Use Cases

Business Process Automation

Development & Testing

Research & Analysis

Architecture

Bytebot is built with:

Why Self-Host?

Advanced Features

Multiple AI Providers

Use any AI provider through our LiteLLM integration:

Enterprise Deployment

Deploy on Kubernetes with Helm:

# Clone the repository
git clone https://github.com/bytebot-ai/bytebot.git
cd bytebot

# Install with Helm
helm install bytebot ./helm \
  --set agent.env.ANTHROPIC_API_KEY=sk-ant-...

Enterprise deployment guide →

Community & Support

Contributing

We welcome contributions! Whether it's:

Please:

  1. Check existing issues first
  2. Open an issue to discuss major changes
  3. Submit PRs with clear descriptions
  4. Join our Discord to discuss ideas

License

Bytebot is open source under the Apache 2.0 license.

Give your AI its own computer. See what it can do.

Deploy on Railway

Built by Tantl Labs and the open source community

bytebot (bytebot-ai/bytebot)

Searching Database…

11.1K ·GitHub
Bytebot
archived
What is this?Bytebot is a self-hosted AI agent that gets its own full Linux desktop (apps, file system, browser) inside a container and completes tasks by actually using that desktop like a human would.
Is this a cloud service or something I run myself?It is fully self-hosted: you run it on your own infrastructure (Docker/Kubernetes) and bring your own AI provider API key, there is no managed Bytebot cloud.
How do I get Bytebot running?Deploy it in about two minutes either with one-click Railway deploy (just add your AI provider API key) or via Docker Compose; there's also a Helm chart for Kubernetes/enterprise deployments.
Which AI providers can I use with it?It works out of the box with Anthropic Claude, OpenAI GPT, and Google Gemini, and through its LiteLLM integration it can also use Azure OpenAI, AWS Bedrock, local Ollama models, and 100+ other providers.
Is this project still maintained?No — the repository is archived, so while the code and self-hosting instructions still work, it is no longer actively developed or updated.
Does my data leave my own servers?No, since Bytebot runs entirely on infrastructure you control, all desktop activity, files, and task data stay on your own servers.
What are the main parts of a Bytebot deployment?Four pieces: a virtual Ubuntu/XFCE desktop with pre-installed apps, an agent service that drives the desktop, a web Task Interface for creating and watching tasks, and REST APIs for programmatic control.
Can it log into websites with my saved passwords?Yes, you can install password managers like 1Password or Bitwarden on its virtual desktop so Bytebot can authenticate automatically, including handling 2FA.
Why use a full desktop agent instead of a browser-only agent?A browser-only agent can't touch native apps, local files, or multi-step workflows across programs; Bytebot's own desktop lets it use VS Code, email clients, office tools, and the file system just like a person.
Does the desktop environment persist between tasks?Yes, programs you install and accounts you log into stay configured on the desktop for future tasks, so you only set things up once.
What are similar projects?sim, Agent-S, evenfire, qwen-code, trueforge, turnstone, magentic-ui, brigade
Which topics is it in?cua, computer-use-agent, desktop-automation, computer-use, ai-tools, nextjs, agentic-ai, gemini, kubernetes, deprecated, anthropic, agents
Creator
bytebot-ai
Size
18.6 MB
Stars
11.1K
Updated
CommitSep 11
Age
1y
Commits · 90 Days