What is Bytebot?
Bytebot is an open-source (Apache 2.0), self-hosted AI desktop agent built by Tantl Labs, a YC-backed startup in San Francisco. It automates computer tasks from natural-language instructions by controlling a full containerized Linux desktop, including the browser, file system, terminal, and code editor, using AI vision rather than site-specific selectors. It can handle logins with 2FA and password managers such as Bitwarden and 1Password. It logs before-and-after screenshots of every action, and a human can step in to take manual control mid-task. It runs via Docker, either locally or on AWS, GCP, or Azure, and works with Claude, OpenAI, Gemini, or LiteLLM as the underlying model. Self-hosting is free beyond the AI provider's API costs and your own infrastructure, with an optional enterprise support tier.
Our verdict
Bytebot fits teams that want an agent operating a real desktop rather than just a browser, and are willing to run and secure their own containers instead of paying for a hosted service. The tradeoff is operational: you own the infrastructure and pay the AI provider's bill yourself. The reported funding figures also come from secondary sources rather than a primary announcement.
Categories:
Agentic automation tools give an LLM control of a browser to navigate, click, and fill forms toward a goal, rather than following a fixed script. They are newer and less predictable than selector-based scraping, but they handle sites and workflows that are hard to hardcode.