Hey everyone, I am the maker of Rasterport. I built this while working with AI agents that needed to interact with frontend apps. A recurring problem was getting reliable access to what a page actually looks like after rendering, in a way that fits cleanly into agent workflows. In many cases, I just wanted something simple. Point to a URL and get back structured browser output that an agent can actually use. Rasterport is my attempt to make that easy. You can run npx rasterport and get a fully rendered screenshot and optional DOM in JSON, ready for agent workflows, automation, or CI. It has already crossed 1100+ downloads in the first 24 hours, which has been a great signal and helped shape early improvements. It works across localhost, staging, and production, and is designed to stay fast and lightweight. This is still v0, and I am actively iterating based on feedback. Would love to hear: - What kind of agent workflows you would use this in - What is missing for your use case - Any rough edges you run into Happy to answer everything here.

Rasterport

Give AI agents eyes on the web

Give AI agents eyes on the web