EDI is one of those things that has never gotten easier. The format is from the 1970s. The tooling feels like it. And the people who have to work with it every day (claims processors, clearinghouses, integration engineers) mostly just suffer through it quietly and get the job done.
We built our own products that depend on healthcare EDI data. We needed a parser that was fast, gave us errors a human could read, and could run anywhere, not just on a server we controlled. We couldn't find one that fit, so we built one. We called it edi-core because the marketing team had other things to worry about.
The thing that makes it different: instead of building a parser, we built something that generates parsers from a single definition. You describe the format once (standard X12 or your own trading partner's companion guide) and it generates a lean, fast parser in the language of your choice. Node, .NET, Python, Java, C++, PHP, Swift. Including in the browser, via WebAssembly.
That last part matters. Your data never leaves your machine. Nothing hits a server. You paste your transaction and get back clean, structured JSON: human-readable field names, code descriptions instead of code values, dates named for what they mean. It looks like data you would have written by hand, because that's what we wanted too.
We're releasing a preview in Josh Labs. Paste an 834 or 837 and see what it can do. If you're building something that needs fast, portable, accurate EDI parsing, we'd love to hear from you.
Try it in Labs.