Hacker News Reader

Does this use schema.org / og meta tags or are you trying to infer object types yourself

It uses a combination of everything to extract information incl. custom parsers. The data returned to the user is always in schema.org.