There is no portable format for a website crawl. Every crawler keeps its results in its own store, so you cannot move a crawl from one tool to another, or into your own systems, without crawling it again. We think that is worth fixing, so Topkay exports your twin in an open format anyone can read.
The format is JSON Graph Format with a small crawl profile. A crawl is a directed graph: each page is a node, each link an edge. The graph carries metadata about the crawl. Each node carries the page's url, status, depth, word count, title, canonical, indexability, internal PageRank, inlinks and when it was last fetched. Each edge carries its link type, anchor text and rel. It reads cleanly in any graph tool and loads into a data warehouse.
Export a verified site three ways from the dashboard, or by posting to the export endpoint with your project key: the JGF graph for interchange, NDJSON with one page per line for BigQuery or Snowflake, and CSV as a floor. See the digital twin and internal PageRank.
The format is versioned and source available. If it is useful, use it. If you build a crawler, emit it, and a crawl becomes something you own and can move, not something locked in a tool. That is the origin thesis applied to your own data.