Bytebot

Bytebot carries out tasks described in plain language on its own virtual desktop. It's for people automating document work, research, or tasks across several applications. The project is archived and no longer maintained.

• Desktop automation: The agent uses browsers and desktop applications, downloads and organizes files, and runs scripts or command-line tools to complete tasks across programs.
• Document processing: You can upload PDFs, spreadsheets, and other files for data extraction, comparisons across documents, or written reports.
• Live viewing and takeover: A web interface lets you watch the desktop as the agent works and take control when a task needs your input.
• Persistent workspace: Installed applications stay available for future tasks, and password managers such as 1Password and Bitwarden can handle sign-ins.
• API access: Other software can create tasks, attach files, and control desktop actions through REST APIs.

Bytebot is open source under the Apache License 2.0. Its Ubuntu desktop runs in Docker on your own infrastructure; Kubernetes deployment with Helm and deployment on Railway are also supported. The desktop includes Firefox and VS Code.

It works with Anthropic Claude, OpenAI GPT, and Google Gemini, plus providers connected through LiteLLM, including local models via Ollama. The desktop and its files live on the infrastructure you choose. Cloud model use requires an internet connection and provider API credentials, and sends task context to that provider.