📣 Understanding differences between schema changes is a little bit easier now!

When lexicon schemas change, it can be difficult to see what exactly is different between the two versions. We now show cleanly formatted diffs to make reviewing changes easy. #ATProtocol #Lexicon

This feature is generally available for anyone to use. There are a few restrictions, however, namely that "dangerous" XRPC methods that involve deleting data or changing critical account information must be explicitly allowed. Some server and admin methods are also outright denied.

📣 Exploring and interacting with lexicons with the help of your favorite agent is a little bit easier now!

We've rolled out llms.txt endpoints across the site, and also released a feature rich MCP that allows you to invoke authenticated XRPC functions.

👀 Beautiful

lexicon.garden/nsid/site.st...

When logged in and viewing your identity, if you have lexicon management enabled then you'll see an edit button next to each of the services listed in your DID document. Use that form to manage the `garden.lexicon.service` records used for this feature.

📣 Using AppViews and developing against them is a little bit easier now!

Lexicon Garden supports service methods and URL templates! This makes it easy to document and broadcast what endpoints you have, what Lexicon methods are provided, and the URLs that can be crafted. #ATProtocol #Lexicon

When you make changes to your lexicon, you can also update your examples. First, you can validate your examples against your lexicon schema to double check that they are still valid. If you need to make a change, you can do so right from the example's view page.

Example records are validated against your schema and give easy to read errors that tell you what needs fixing and where. This takes the guess-work out of creating easy to use documentation and example content.

📣 Creating lexicon schema examples is a little bit easier now!

When viewing a lexicon schema that you own, you can create examples of the lexicon with just two clicks. Visit the lexicon schema page and jump to the example section. #ATProtocol #Documentation #Lexicon

📣 Validating procedures and queries is just a little bit easier now!

From a lexicon page, you can validate XRPC procedure and query definitions!

Under the hood this uses the same /xrpc/com.atproto.lexicon.validate endpoint as lexicon.garden/validate

📣 Validating lexicons is just a little bit easier now!

From a lexicon page, you can validate the lexicon schema with or without its dependencies, as well as user-supplied records.

Under the hood this uses the same /xrpc/com.atproto.lexicon.validate endpoint as lexicon.garden/validate

Here's an example in action:

lexicon.garden/example/did:...

✅ Cleanly formatted records
✅ Validation checks
✅ Referenced directly in documentation
✅ Easy to find and navigate to

The graph is populated by direct Lexicon schema relationships and also developer examples. The com.atproto.repo.strongRef type is powerful, but not very informative. Example records with references populate the graph as well, like with the calendar RSVP graph above.

lexicon.garden/example/did:...

This works for query, procedure, and subscription XRPC methods too! You can see what types a method returns or accepts as input.

lexicon.garden/nsid/app.bsky.feed.getPostThread

Graphs also show permission sets. If a lexicon defines which record types an OAuth scope can access, those relationships appear in the graph too.

lexicon.garden/nsid/dev.baileytownsend.demo.authBasicFeatures

Others start simple but open up. `community.lexicon.calendar.rsvp` has one edge to the event type, but that node connects to location schemas, geo coordinates, and more.

lexicon.garden/nsid/community.lexicon.calendar.rsvp

Some graphs are dense. `app.bsky.feed.post` connects to over a dozen types: embed formats, facets, labels, reply references. You can see how one record type branches out into the whole embed system.

lexicon.garden/nsid/app.bsky.feed.post