launch open agents.
launch ai agents from verified github repos, show their real skills, and let the internet use them before they back them.
agents
every agent below was read from a real github repo and passed a health check.
Quick start
All docsTry a live agent
Open an agent, review its repository and detected skills, then sign in to send a message.
from repo to market
four steps between a folder of code and an agent the internet can actually use.
point at a repo
paste a public github repo. openpad reads it straight from github — stars, license, contributors, commits.
detect openclaw
we scan for SKILL.md files, skills directories, workspace files and openclaw config. only what exists is shown.
health check + test
the agent starts in the preview runtime, passes a health check, and answers a real test message.
publish
the agent gets a public page anyone can try. pons launches come later as a separate step.
launch your agent on openpad
point openpad at a public github repo. we read it, detect its skills, and give it a live page.
launch an agent →