diff --git a/examples/json-schema-to-grammar.py b/examples/json-schema-to-grammar.py
index ce090dd59..bedeb18d1 100755
--- a/examples/json-schema-to-grammar.py
+++ b/examples/json-schema-to-grammar.py
@@ -33,7 +33,6 @@ class SchemaConverter:
     def __init__(self, prop_order):
         self._prop_order = prop_order
         self._rules = {'space': SPACE_RULE}
-        self.ref_base = None
 
     def _format_literal(self, literal):
         escaped = GRAMMAR_LITERAL_ESCAPE_RE.sub(
@@ -53,20 +52,44 @@ class SchemaConverter:
         self._rules[key] = rule
         return key
 
-    def _resolve_ref(self, ref):
-        # TODO: use https://github.com/APIDevTools/json-schema-ref-parser
-        try:
-            if ref is not None and ref.startswith('#/'):# and 'definitions' in schema:
-                target = self.ref_base
-                name = None
-                for sel in ref.split('/')[1:]:
-                    name = sel
-                    target = target[sel]
-                return (name, target)
-            return None
-        except KeyError as e:
-            raise Exception(f'Error resolving ref {ref}: {e}') from e
-    
+    @staticmethod
+    def resolve_refs(schema: dict, url_cache: Dict[str, dict] = None):
+        if url_cache is None:
+            url_cache = {}
+
+        def visit(n: dict):
+            if isinstance(n, list):
+                return [visit(x) for x in n]
+            elif isinstance(n, dict):
+                ref = n.get('$ref')
+                if ref is not None:
+                    if ref.startswith('#/'):
+                        target = schema
+                        for sel in ref.split('/')[1:]:
+                            assert target is not None and sel in target, f'Error resolving ref {ref}: {sel} not in {target}'
+                            target = target[sel]
+                        
+                        return target
+                    elif ref.startswith('https://'):
+                        import requests
+                        linked_schema = url_cache.get(ref)
+                        if linked_schema is None:
+                            linked_schema = requests.get(ref).json()
+                            url_cache[ref] = linked_schema
+                        return SchemaConverter.resolve_refs(linked_schema, url_cache)
+                    else:
+                        raise ValueError(f'Unsupported ref {ref}')
+                else:
+                    for k in n.keys():
+                        v = n[k]
+                        vv = visit(v)
+                        if vv is not v:
+                            n[k] = vv
+            else:
+                pass
+            return n
+        return visit(schema)
+
     def _generate_union_rule(self, name, alt_schemas):
         return ' | '.join((
             self.visit(alt_schema, f'{name}{"-" if name else ""}{i}')
@@ -163,17 +186,8 @@ class SchemaConverter:
             raise Exception(f'Error processing pattern: {pattern}: {e}') from e
 
     def visit(self, schema, name):
-        old_ref_base = self.ref_base
-        if 'definitions' in schema:
-            self.ref_base = schema
-        try:
-            return self._visit(schema, name)
-        finally:
-            self.ref_base = old_ref_base
-    
-    def _visit(self, schema, name):
+        assert '$ref' not in schema, f'Unresolved $ref in {schema} (make sure to use {SchemaConverter.resolve_refs.__name__})'
         schema_type = schema.get('type')
-        ref = schema.get('$ref')
         rule_name = name or 'root'
 
         if 'oneOf' in schema or 'anyOf' in schema:
@@ -198,9 +212,6 @@ class SchemaConverter:
             required = set()
             properties = []
             def add_component(comp_schema, is_required):
-                ref = comp_schema.get('$ref')
-                if ref is not None and (resolved := self._resolve_ref(ref)) is not None:
-                    comp_schema = resolved[1]
 
                 if 'properties' in comp_schema:
                     for prop_name, prop_schema in comp_schema['properties'].items():
@@ -262,16 +273,6 @@ class SchemaConverter:
         elif schema_type in (None, 'string') and 'pattern' in schema:
             return self._add_rule(rule_name, self._visit_pattern(schema['pattern'], rule_name))
 
-        elif (resolved := self._resolve_ref(ref)) is not None:
-            (ref_name, definition) = resolved
-            def_name = f'{name}-{ref_name}' if name else ''
-            return self.visit(definition, def_name)
-        
-        elif ref is not None and ref.startswith('https://'):
-            import requests
-            ref_schema = requests.get(ref).json()
-            return self.visit(ref_schema, ref)
-
         elif schema_type == 'object' and len(schema) == 1 or schema_type is None and len(schema) == 0:
             # return 'object'
             for t, r in PRIMITIVE_RULES.items():
@@ -286,9 +287,7 @@ class SchemaConverter:
             )
     
     def _build_object_rule(self, properties: List[Tuple[str, Any]], required: Set[str], name: str):
-        # TODO: `required` keyword
         prop_order = self._prop_order
-        print(f'# properties: {properties}', file=sys.stderr)
         # sort by position in prop_order (if specified) then by original order
         sorted_props = [kv[0] for _, kv in sorted(enumerate(properties), key=lambda ikv: (prop_order.get(ikv[1][0], len(prop_order)), ikv[0]))]
 
@@ -370,6 +369,7 @@ def main(args_in = None):
     else:
         with open(args.schema) as f:
             schema = json.load(f)
+    schema = SchemaConverter.resolve_refs(schema)
     prop_order = {name: idx for idx, name in enumerate(args.prop_order)}
     converter = SchemaConverter(prop_order)
     converter.visit(schema, '')