Sediman is an open-source AI agent that automates browser workflows by learning from your actions. Unlike RPA tools that break on UI changes, Sediman understands context — watch once, repeat forever. Watch & Learn, Skills system (shareable automation blocks like an App Store), Open-source core in Python + Rust + TypeScript, self-hosted, no cloud dependency. Built for growth teams doing repetitive browser work — scraping, posting, lead research. Anyone paying a VA to click buttons.
No reviews yetBe the first to leave a review for Sediman Agent
Maker
📌
Hey everyone! 👋 Jason here, solo dev behind Sediman.
I built this because I was tired of breaking Selenium scripts every time a website changed a class name. Sediman takes a different approach — it understands what you're doing, not just where you're clicking.
What browser task would you automate first? I'd love to hear what repetitive work eats your time — we might already have a skill for it.
Happy to answer any technical questions about the architecture. We've got a Rust TUI, Python agent core with sandboxed subprocess execution, and a 3-language plugin system. It's a beast.