diff --git a/prisma-fmt/src/lib.rs b/prisma-fmt/src/lib.rs
index 256b5744887b..f2838915bcbb 100644
--- a/prisma-fmt/src/lib.rs
+++ b/prisma-fmt/src/lib.rs
@@ -7,6 +7,7 @@ mod merge_schemas;
 mod native;
 mod offsets;
 mod preview;
+mod references;
 mod schema_file_input;
 mod text_document_completion;
 mod validate;
@@ -84,13 +85,30 @@ pub fn code_actions(schema_files: String, params: &str) -> String {
 
     let Ok(input) = serde_json::from_str::<SchemaFileInput>(&schema_files) else {
         warn!("Failed to parse schema file input");
-        return serde_json::to_string(&text_document_completion::empty_completion_list()).unwrap();
+        return serde_json::to_string(&code_actions::empty_code_actions()).unwrap();
     };
 
     let actions = code_actions::available_actions(input.into(), params);
     serde_json::to_string(&actions).unwrap()
 }
 
+pub fn references(schema_files: String, params: &str) -> String {
+    let params: lsp_types::ReferenceParams = if let Ok(params) = serde_json::from_str(params) {
+        params
+    } else {
+        warn!("Failed to parse params to references() as ReferenceParams.");
+        return serde_json::to_string(&references::empty_references()).unwrap();
+    };
+
+    let Ok(input) = serde_json::from_str::<SchemaFileInput>(&schema_files) else {
+        warn!("Failed to parse schema file input");
+        return serde_json::to_string(&references::empty_references()).unwrap();
+    };
+
+    let references = references::references(input.into(), params);
+    serde_json::to_string(&references).unwrap()
+}
+
 /// The two parameters are:
 /// - The [`SchemaFileInput`] to reformat, as a string.
 /// - An LSP
diff --git a/prisma-fmt/src/references.rs b/prisma-fmt/src/references.rs
new file mode 100644
index 000000000000..00da04548893
--- /dev/null
+++ b/prisma-fmt/src/references.rs
@@ -0,0 +1,245 @@
+use log::*;
+use lsp_types::{Location, ReferenceParams, Url};
+use psl::{
+    diagnostics::FileId,
+    error_tolerant_parse_configuration,
+    parser_database::ParserDatabase,
+    schema_ast::ast::{
+        AttributePosition, CompositeTypePosition, EnumPosition, Field, FieldId, FieldPosition, FieldType, Identifier,
+        ModelId, ModelPosition, SchemaPosition, SourcePosition, Top, WithIdentifier, WithName,
+    },
+    Diagnostics, SourceFile,
+};
+
+use crate::{offsets::position_to_offset, span_to_range, LSPContext};
+
+pub(super) type ReferencesContext<'a> = LSPContext<'a, ReferenceParams>;
+
+pub(crate) fn empty_references() -> Vec<Location> {
+    Vec::new()
+}
+
+fn empty_identifiers<'ast>() -> impl Iterator<Item = &'ast Identifier> {
+    std::iter::empty()
+}
+
+pub(crate) fn references(schema_files: Vec<(String, SourceFile)>, params: ReferenceParams) -> Vec<Location> {
+    let (_, config, _) = error_tolerant_parse_configuration(&schema_files);
+
+    let db = {
+        let mut diag = Diagnostics::new();
+        ParserDatabase::new(&schema_files, &mut diag)
+    };
+
+    let Some(initiating_file_id) = db.file_id(params.text_document_position.text_document.uri.as_str()) else {
+        warn!("Initating file name is not found in the schema");
+        return empty_references();
+    };
+
+    let initiating_doc = db.source(initiating_file_id);
+
+    let position = if let Some(pos) = position_to_offset(&params.text_document_position.position, initiating_doc) {
+        pos
+    } else {
+        warn!("Received a position outside of the document boundaries in ReferenceParams");
+        return empty_references();
+    };
+
+    let target_position = db.ast(initiating_file_id).find_at_position(position);
+
+    let ctx = ReferencesContext {
+        db: &db,
+        config: &config,
+        initiating_file_id,
+        params: &params,
+    };
+
+    reference_locations_for_target(ctx, target_position)
+}
+
+fn reference_locations_for_target(ctx: ReferencesContext<'_>, target: SchemaPosition) -> Vec<Location> {
+    let identifiers: Vec<&Identifier> = match target {
+        // Blocks
+        SchemaPosition::Model(model_id, ModelPosition::Name(name)) => {
+            let model = ctx.db.walk((ctx.initiating_file_id, model_id));
+
+            std::iter::once(model.ast_model().identifier())
+                .chain(find_where_used_as_field_type(&ctx, name))
+                .collect()
+        }
+        SchemaPosition::Enum(enum_id, EnumPosition::Name(name)) => {
+            let enm = ctx.db.walk((ctx.initiating_file_id, enum_id));
+
+            std::iter::once(enm.ast_enum().identifier())
+                .chain(find_where_used_as_field_type(&ctx, name))
+                .collect()
+        }
+        SchemaPosition::CompositeType(composite_id, CompositeTypePosition::Name(name)) => {
+            let ct = ctx.db.walk((ctx.initiating_file_id, composite_id));
+
+            std::iter::once(ct.ast_composite_type().identifier())
+                .chain(find_where_used_as_field_type(&ctx, name))
+                .collect()
+        }
+        SchemaPosition::DataSource(_, SourcePosition::Name(name)) => find_where_used_as_ds_name(&ctx, name)
+            .into_iter()
+            .chain(find_where_used_for_native_type(&ctx, name))
+            .collect(),
+
+        // Fields
+        SchemaPosition::Model(_, ModelPosition::Field(_, FieldPosition::Type(r#type)))
+        | SchemaPosition::CompositeType(_, CompositeTypePosition::Field(_, FieldPosition::Type(r#type))) => {
+            find_where_used_as_top_name(&ctx, r#type)
+                .into_iter()
+                .chain(find_where_used_as_field_type(&ctx, r#type))
+                .collect()
+        }
+
+        // Attributes
+        SchemaPosition::Model(
+            model_id,
+            ModelPosition::Field(
+                field_id,
+                FieldPosition::Attribute(_, _, AttributePosition::ArgumentValue(arg_name, arg_value)),
+            ),
+        ) => match arg_name {
+            Some("fields") => find_where_used_as_field_name(&ctx, arg_value.as_str(), model_id, ctx.initiating_file_id)
+                .into_iter()
+                .collect(),
+            Some("references") => {
+                let field = &ctx.db.ast(ctx.initiating_file_id)[model_id][field_id];
+                let referenced_model = field.field_type.name();
+
+                let Some(ref_model_id) = ctx.db.find_model(referenced_model) else {
+                    warn!("Could not find model with name: {}", referenced_model);
+                    return empty_references();
+                };
+
+                find_where_used_as_field_name(&ctx, arg_value.as_str(), ref_model_id.id.1, ref_model_id.id.0)
+                    .into_iter()
+                    .collect()
+            }
+            _ => vec![],
+        },
+
+        // ? This might make more sense to add as a definition rather than a reference
+        SchemaPosition::Model(_, ModelPosition::Field(_, FieldPosition::Attribute(name, _, _)))
+        | SchemaPosition::CompositeType(_, CompositeTypePosition::Field(_, FieldPosition::Attribute(name, _, _))) => {
+            match ctx.datasource().map(|ds| &ds.name) {
+                Some(ds_name) if name.contains(ds_name) => find_where_used_as_ds_name(&ctx, ds_name)
+                    .into_iter()
+                    .chain(find_where_used_for_native_type(&ctx, ds_name))
+                    .collect(),
+                _ => vec![],
+            }
+        }
+
+        SchemaPosition::Model(
+            model_id,
+            ModelPosition::ModelAttribute(_attr_name, _, AttributePosition::ArgumentValue(_, arg_val)),
+        ) => find_where_used_as_field_name(&ctx, arg_val.as_str(), model_id, ctx.initiating_file_id)
+            .into_iter()
+            .collect(),
+
+        _ => vec![],
+    };
+
+    identifiers
+        .iter()
+        .filter_map(|ident| ident_to_location(ident, &ctx))
+        .collect()
+}
+
+fn find_where_used_as_field_name<'ast>(
+    ctx: &'ast ReferencesContext<'_>,
+    name: &str,
+    model_id: ModelId,
+    file_id: FileId,
+) -> Option<&'ast Identifier> {
+    let model = ctx.db.walk((file_id, model_id));
+
+    match model.scalar_fields().find(|field| field.name() == name) {
+        Some(field) => Some(field.ast_field().identifier()),
+        None => None,
+    }
+}
+
+fn find_where_used_for_native_type<'ast>(
+    ctx: &ReferencesContext<'ast>,
+    name: &'ast str,
+) -> impl Iterator<Item = &'ast Identifier> {
+    fn find_native_type_locations<'ast>(
+        name: &'ast str,
+        fields: impl Iterator<Item = (FieldId, &'ast Field)> + 'ast,
+    ) -> Box<dyn Iterator<Item = &'ast Identifier> + 'ast> {
+        Box::new(fields.filter_map(move |field| {
+            field
+                .1
+                .attributes
+                .iter()
+                .find(|attr| extract_ds_from_native_type(attr.name()) == Some(name))
+                .map(|attr| attr.identifier())
+        }))
+    }
+
+    ctx.db.walk_tops().flat_map(move |top| match top.ast_top() {
+        Top::CompositeType(composite_type) => find_native_type_locations(name, composite_type.iter_fields()),
+        Top::Model(model) => find_native_type_locations(name, model.iter_fields()),
+
+        Top::Enum(_) | Top::Source(_) | Top::Generator(_) => Box::new(empty_identifiers()),
+    })
+}
+
+fn find_where_used_as_field_type<'ast>(
+    ctx: &'ast ReferencesContext<'_>,
+    name: &'ast str,
+) -> impl Iterator<Item = &'ast Identifier> {
+    fn get_relevent_identifiers<'a>(
+        fields: impl Iterator<Item = (FieldId, &'a Field)>,
+        name: &str,
+    ) -> Vec<&'a Identifier> {
+        fields
+            .filter_map(|(_id, field)| match &field.field_type {
+                FieldType::Supported(id) if id.name == name => Some(id),
+                _ => None,
+            })
+            .collect()
+    }
+
+    ctx.db.walk_tops().flat_map(|top| match top.ast_top() {
+        Top::Model(model) => get_relevent_identifiers(model.iter_fields(), name),
+        Top::CompositeType(composite_type) => get_relevent_identifiers(composite_type.iter_fields(), name),
+        // * Cannot contain field types
+        Top::Enum(_) | Top::Source(_) | Top::Generator(_) => vec![],
+    })
+}
+
+fn find_where_used_as_top_name<'ast>(ctx: &'ast ReferencesContext<'_>, name: &'ast str) -> Option<&'ast Identifier> {
+    ctx.db.find_top(name).map(|top| top.ast_top().identifier())
+}
+
+fn find_where_used_as_ds_name<'ast>(ctx: &'ast ReferencesContext<'_>, name: &'ast str) -> Option<&'ast Identifier> {
+    ctx.db
+        .find_source(name)
+        .map(|source| ctx.db.ast(source.0)[source.1].identifier())
+}
+
+fn extract_ds_from_native_type(attr_name: &str) -> Option<&str> {
+    attr_name.split('.').next()
+}
+
+fn ident_to_location<'ast>(id: &'ast Identifier, ctx: &'ast ReferencesContext<'_>) -> Option<Location> {
+    let file_id = id.span.file_id;
+
+    let source = ctx.db.source(file_id);
+    let range = span_to_range(id.span, source);
+    let file_name = ctx.db.file_name(file_id);
+
+    let uri = if let Ok(uri) = Url::parse(file_name) {
+        uri
+    } else {
+        return None;
+    };
+
+    Some(Location { uri, range })
+}
diff --git a/prisma-fmt/src/text_document_completion.rs b/prisma-fmt/src/text_document_completion.rs
index d1234c80b40c..b791a69c3515 100644
--- a/prisma-fmt/src/text_document_completion.rs
+++ b/prisma-fmt/src/text_document_completion.rs
@@ -5,7 +5,8 @@ use lsp_types::*;
 use psl::{
     diagnostics::Span,
     error_tolerant_parse_configuration,
-    parser_database::{ast, ParserDatabase, SourceFile},
+    parser_database::{ast, ParserDatabase, ReferentialAction, SourceFile},
+    schema_ast::ast::AttributePosition,
     Diagnostics, PreviewFeature,
 };
 
@@ -85,19 +86,46 @@ fn push_ast_completions(ctx: CompletionContext<'_>, completion_list: &mut Comple
         .relation_mode()
         .unwrap_or_else(|| ctx.connector().default_relation_mode());
 
-    match ctx.db.ast(ctx.initiating_file_id).find_at_position(position) {
+    let find_at_position = ctx.db.ast(ctx.initiating_file_id).find_at_position(position);
+
+    fn push_referential_action(completion_list: &mut CompletionList, referential_action: ReferentialAction) {
+        completion_list.items.push(CompletionItem {
+            label: referential_action.as_str().to_owned(),
+            kind: Some(CompletionItemKind::ENUM),
+            // what is the difference between detail and documentation?
+            detail: Some(referential_action.documentation().to_owned()),
+            ..Default::default()
+        });
+    }
+
+    match find_at_position {
         ast::SchemaPosition::Model(
             _model_id,
-            ast::ModelPosition::Field(_, ast::FieldPosition::Attribute("relation", _, Some(attr_name))),
+            ast::ModelPosition::Field(
+                _,
+                ast::FieldPosition::Attribute("relation", _, AttributePosition::Argument(attr_name)),
+            ),
         ) if attr_name == "onDelete" || attr_name == "onUpdate" => {
             for referential_action in ctx.connector().referential_actions(&relation_mode).iter() {
-                completion_list.items.push(CompletionItem {
-                    label: referential_action.as_str().to_owned(),
-                    kind: Some(CompletionItemKind::ENUM),
-                    // what is the difference between detail and documentation?
-                    detail: Some(referential_action.documentation().to_owned()),
-                    ..Default::default()
-                });
+                push_referential_action(completion_list, referential_action);
+            }
+        }
+
+        ast::SchemaPosition::Model(
+            _model_id,
+            ast::ModelPosition::Field(
+                _,
+                ast::FieldPosition::Attribute("relation", _, AttributePosition::ArgumentValue(attr_name, value)),
+            ),
+        ) => {
+            if let Some(attr_name) = attr_name {
+                if attr_name == "onDelete" || attr_name == "onUpdate" {
+                    ctx.connector()
+                        .referential_actions(&relation_mode)
+                        .iter()
+                        .filter(|ref_action| ref_action.to_string().starts_with(&value))
+                        .for_each(|ref_action| push_referential_action(completion_list, ref_action));
+                }
             }
         }
 
diff --git a/prisma-fmt/tests/references/mod.rs b/prisma-fmt/tests/references/mod.rs
new file mode 100644
index 000000000000..cf3a59fec326
--- /dev/null
+++ b/prisma-fmt/tests/references/mod.rs
@@ -0,0 +1,2 @@
+mod test_api;
+mod tests;
diff --git a/prisma-fmt/tests/references/scenarios/composite_type_as_type/a.prisma b/prisma-fmt/tests/references/scenarios/composite_type_as_type/a.prisma
new file mode 100644
index 000000000000..0c4f8cbe2e10
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/composite_type_as_type/a.prisma
@@ -0,0 +1,14 @@
+generator client {
+  provider        = "prisma-client-js"
+  previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+  provider = "mongodb"
+  url      = env("DATABASE_URL")
+}
+
+type Address {
+  city     String
+  postCode String
+}
diff --git a/prisma-fmt/tests/references/scenarios/composite_type_as_type/b.prisma b/prisma-fmt/tests/references/scenarios/composite_type_as_type/b.prisma
new file mode 100644
index 000000000000..92839a1f9ac3
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/composite_type_as_type/b.prisma
@@ -0,0 +1,5 @@
+model User {
+  id       String @id @map("_id")
+  authorId String
+  address  Add<|>ress
+}
diff --git a/prisma-fmt/tests/references/scenarios/composite_type_as_type/result.json b/prisma-fmt/tests/references/scenarios/composite_type_as_type/result.json
new file mode 100644
index 000000000000..d02299ab411e
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/composite_type_as_type/result.json
@@ -0,0 +1,28 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 10,
+        "character": 5
+      },
+      "end": {
+        "line": 10,
+        "character": 12
+      }
+    }
+  },
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 3,
+        "character": 11
+      },
+      "end": {
+        "line": 3,
+        "character": 18
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/composite_type_name/a.prisma b/prisma-fmt/tests/references/scenarios/composite_type_name/a.prisma
new file mode 100644
index 000000000000..30df23f42ecb
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/composite_type_name/a.prisma
@@ -0,0 +1,14 @@
+generator client {
+  provider        = "prisma-client-js"
+  previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+  provider = "mongodb"
+  url      = env("DATABASE_URL")
+}
+
+type Add<|>ress {
+city     String
+postCode String
+}
diff --git a/prisma-fmt/tests/references/scenarios/composite_type_name/b.prisma b/prisma-fmt/tests/references/scenarios/composite_type_name/b.prisma
new file mode 100644
index 000000000000..9864e134aaeb
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/composite_type_name/b.prisma
@@ -0,0 +1,5 @@
+model User {
+  id       String  @id @map("_id")
+  authorId String
+  address  Address
+}
diff --git a/prisma-fmt/tests/references/scenarios/composite_type_name/result.json b/prisma-fmt/tests/references/scenarios/composite_type_name/result.json
new file mode 100644
index 000000000000..d02299ab411e
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/composite_type_name/result.json
@@ -0,0 +1,28 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 10,
+        "character": 5
+      },
+      "end": {
+        "line": 10,
+        "character": 12
+      }
+    }
+  },
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 3,
+        "character": 11
+      },
+      "end": {
+        "line": 3,
+        "character": 18
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/datasource_as_attribute/a.prisma b/prisma-fmt/tests/references/scenarios/datasource_as_attribute/a.prisma
new file mode 100644
index 000000000000..a48d0d496a2f
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/datasource_as_attribute/a.prisma
@@ -0,0 +1,15 @@
+generator client {
+  provider        = "prisma-client-js"
+  previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+  provider = "mongodb"
+  url      = env("DATABASE_URL")
+}
+
+enum Pet {
+  Cat
+  Dog
+  Redpanda
+}
diff --git a/prisma-fmt/tests/references/scenarios/datasource_as_attribute/b.prisma b/prisma-fmt/tests/references/scenarios/datasource_as_attribute/b.prisma
new file mode 100644
index 000000000000..8406f54dbd58
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/datasource_as_attribute/b.prisma
@@ -0,0 +1,4 @@
+model User {
+  id  String @id @map("_id") @d<|>b.String
+  pet Pet
+}
diff --git a/prisma-fmt/tests/references/scenarios/datasource_as_attribute/result.json b/prisma-fmt/tests/references/scenarios/datasource_as_attribute/result.json
new file mode 100644
index 000000000000..d9905f2e7ade
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/datasource_as_attribute/result.json
@@ -0,0 +1,28 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 5,
+        "character": 11
+      },
+      "end": {
+        "line": 5,
+        "character": 13
+      }
+    }
+  },
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 1,
+        "character": 30
+      },
+      "end": {
+        "line": 1,
+        "character": 39
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/datasource_name/a.prisma b/prisma-fmt/tests/references/scenarios/datasource_name/a.prisma
new file mode 100644
index 000000000000..201b7bd2c2cd
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/datasource_name/a.prisma
@@ -0,0 +1,15 @@
+generator client {
+  provider        = "prisma-client-js"
+  previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource d<|>b {
+provider = "mongodb"
+url      = env("DATABASE_URL")
+}
+
+enum Pet {
+  Cat
+  Dog
+  Redpanda
+}
diff --git a/prisma-fmt/tests/references/scenarios/datasource_name/b.prisma b/prisma-fmt/tests/references/scenarios/datasource_name/b.prisma
new file mode 100644
index 000000000000..16d0c16fe768
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/datasource_name/b.prisma
@@ -0,0 +1,4 @@
+model User {
+  id  String @id @map("_id") @db.String
+  pet Pet
+}
diff --git a/prisma-fmt/tests/references/scenarios/datasource_name/result.json b/prisma-fmt/tests/references/scenarios/datasource_name/result.json
new file mode 100644
index 000000000000..d9905f2e7ade
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/datasource_name/result.json
@@ -0,0 +1,28 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 5,
+        "character": 11
+      },
+      "end": {
+        "line": 5,
+        "character": 13
+      }
+    }
+  },
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 1,
+        "character": 30
+      },
+      "end": {
+        "line": 1,
+        "character": 39
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/enum_as_type/a.prisma b/prisma-fmt/tests/references/scenarios/enum_as_type/a.prisma
new file mode 100644
index 000000000000..a48d0d496a2f
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/enum_as_type/a.prisma
@@ -0,0 +1,15 @@
+generator client {
+  provider        = "prisma-client-js"
+  previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+  provider = "mongodb"
+  url      = env("DATABASE_URL")
+}
+
+enum Pet {
+  Cat
+  Dog
+  Redpanda
+}
diff --git a/prisma-fmt/tests/references/scenarios/enum_as_type/b.prisma b/prisma-fmt/tests/references/scenarios/enum_as_type/b.prisma
new file mode 100644
index 000000000000..a8b01dd0dce0
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/enum_as_type/b.prisma
@@ -0,0 +1,4 @@
+model User {
+  id String @id @map("_id")
+  pet P<|>et
+}
diff --git a/prisma-fmt/tests/references/scenarios/enum_as_type/result.json b/prisma-fmt/tests/references/scenarios/enum_as_type/result.json
new file mode 100644
index 000000000000..8c2123cedda6
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/enum_as_type/result.json
@@ -0,0 +1,28 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 10,
+        "character": 5
+      },
+      "end": {
+        "line": 10,
+        "character": 8
+      }
+    }
+  },
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 2,
+        "character": 6
+      },
+      "end": {
+        "line": 2,
+        "character": 9
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/enum_name/a.prisma b/prisma-fmt/tests/references/scenarios/enum_name/a.prisma
new file mode 100644
index 000000000000..d227fc11774b
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/enum_name/a.prisma
@@ -0,0 +1,15 @@
+generator client {
+  provider        = "prisma-client-js"
+  previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+  provider = "mongodb"
+  url      = env("DATABASE_URL")
+}
+
+enum P<|>et {
+Cat
+Dog
+Redpanda
+}
diff --git a/prisma-fmt/tests/references/scenarios/enum_name/b.prisma b/prisma-fmt/tests/references/scenarios/enum_name/b.prisma
new file mode 100644
index 000000000000..0e2191c72b30
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/enum_name/b.prisma
@@ -0,0 +1,4 @@
+model User {
+  id  String @id @map("_id")
+  pet Pet
+}
diff --git a/prisma-fmt/tests/references/scenarios/enum_name/result.json b/prisma-fmt/tests/references/scenarios/enum_name/result.json
new file mode 100644
index 000000000000..8c2123cedda6
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/enum_name/result.json
@@ -0,0 +1,28 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 10,
+        "character": 5
+      },
+      "end": {
+        "line": 10,
+        "character": 8
+      }
+    }
+  },
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 2,
+        "character": 6
+      },
+      "end": {
+        "line": 2,
+        "character": 9
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/model_as_type/a.prisma b/prisma-fmt/tests/references/scenarios/model_as_type/a.prisma
new file mode 100644
index 000000000000..e050e2e32715
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_as_type/a.prisma
@@ -0,0 +1,16 @@
+generator client {
+  provider        = "prisma-client-js"
+  previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+  provider = "mongodb"
+  url      = env("DATABASE_URL")
+}
+
+model UserTwo {
+  id String @id @map("_id") @db.String
+
+  posts  Post   @relation(fields: [postId], references: [id])
+  postId String
+}
diff --git a/prisma-fmt/tests/references/scenarios/model_as_type/b.prisma b/prisma-fmt/tests/references/scenarios/model_as_type/b.prisma
new file mode 100644
index 000000000000..58f0d86dae09
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_as_type/b.prisma
@@ -0,0 +1,5 @@
+model Post {
+  id       String    @id @map("_id")
+  authorId String
+  UserTwo  User<|>Two[]
+}
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/model_as_type/result.json b/prisma-fmt/tests/references/scenarios/model_as_type/result.json
new file mode 100644
index 000000000000..3e31baf3b169
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_as_type/result.json
@@ -0,0 +1,28 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 10,
+        "character": 6
+      },
+      "end": {
+        "line": 10,
+        "character": 13
+      }
+    }
+  },
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 3,
+        "character": 11
+      },
+      "end": {
+        "line": 3,
+        "character": 18
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/model_name/a.prisma b/prisma-fmt/tests/references/scenarios/model_name/a.prisma
new file mode 100644
index 000000000000..8e723bdc9e5c
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_name/a.prisma
@@ -0,0 +1,16 @@
+generator client {
+  provider        = "prisma-client-js"
+  previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+  provider = "mongodb"
+  url      = env("DATABASE_URL")
+}
+
+model User<|>Two {
+id String @id @map("_id") @db.String
+
+posts  Post   @relation(fields: [postId], references: [id])
+postId String
+}
diff --git a/prisma-fmt/tests/references/scenarios/model_name/b.prisma b/prisma-fmt/tests/references/scenarios/model_name/b.prisma
new file mode 100644
index 000000000000..107a84502a96
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_name/b.prisma
@@ -0,0 +1,5 @@
+model Post {
+  id       String    @id @map("_id")
+  authorId String
+  UserTwo  UserTwo[]
+}
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/model_name/result.json b/prisma-fmt/tests/references/scenarios/model_name/result.json
new file mode 100644
index 000000000000..3e31baf3b169
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_name/result.json
@@ -0,0 +1,28 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 10,
+        "character": 6
+      },
+      "end": {
+        "line": 10,
+        "character": 13
+      }
+    }
+  },
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 3,
+        "character": 11
+      },
+      "end": {
+        "line": 3,
+        "character": 18
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/model_relation_fields/a.prisma b/prisma-fmt/tests/references/scenarios/model_relation_fields/a.prisma
new file mode 100644
index 000000000000..835f01b07f01
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_relation_fields/a.prisma
@@ -0,0 +1,7 @@
+model Compound {
+    id      String
+    name    String
+    UserTwo UserTwo[]
+
+    @@id([id, name])
+}
diff --git a/prisma-fmt/tests/references/scenarios/model_relation_fields/b.prisma b/prisma-fmt/tests/references/scenarios/model_relation_fields/b.prisma
new file mode 100644
index 000000000000..7308f18cd5b3
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_relation_fields/b.prisma
@@ -0,0 +1,7 @@
+model UserTwo {
+    id String @id @map("_id")
+
+    compounds    Compound @relation(fields: [comp<|>oundId, compoundName], references: [id, name])
+    compoundId   String
+    compoundName String
+}
diff --git a/prisma-fmt/tests/references/scenarios/model_relation_fields/config.prisma b/prisma-fmt/tests/references/scenarios/model_relation_fields/config.prisma
new file mode 100644
index 000000000000..523204f6f2f5
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_relation_fields/config.prisma
@@ -0,0 +1,9 @@
+generator client {
+    provider        = "prisma-client-js"
+    previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+    provider = "postgres"
+    url      = env("DATABASE_URL")
+}
diff --git a/prisma-fmt/tests/references/scenarios/model_relation_fields/result.json b/prisma-fmt/tests/references/scenarios/model_relation_fields/result.json
new file mode 100644
index 000000000000..fbbe001172cf
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_relation_fields/result.json
@@ -0,0 +1,15 @@
+[
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 4,
+        "character": 4
+      },
+      "end": {
+        "line": 4,
+        "character": 14
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/model_relation_references/a.prisma b/prisma-fmt/tests/references/scenarios/model_relation_references/a.prisma
new file mode 100644
index 000000000000..835f01b07f01
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_relation_references/a.prisma
@@ -0,0 +1,7 @@
+model Compound {
+    id      String
+    name    String
+    UserTwo UserTwo[]
+
+    @@id([id, name])
+}
diff --git a/prisma-fmt/tests/references/scenarios/model_relation_references/b.prisma b/prisma-fmt/tests/references/scenarios/model_relation_references/b.prisma
new file mode 100644
index 000000000000..ff5c52bdc6d1
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_relation_references/b.prisma
@@ -0,0 +1,7 @@
+model UserTwo {
+    id String @id @map("_id")
+
+    compounds    Compound @relation(fields: [compoundId, compoundName], references: [id, n<|>ame])
+    compoundId   String
+    compoundName String
+}
diff --git a/prisma-fmt/tests/references/scenarios/model_relation_references/config.prisma b/prisma-fmt/tests/references/scenarios/model_relation_references/config.prisma
new file mode 100644
index 000000000000..523204f6f2f5
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_relation_references/config.prisma
@@ -0,0 +1,9 @@
+generator client {
+    provider        = "prisma-client-js"
+    previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+    provider = "postgres"
+    url      = env("DATABASE_URL")
+}
diff --git a/prisma-fmt/tests/references/scenarios/model_relation_references/result.json b/prisma-fmt/tests/references/scenarios/model_relation_references/result.json
new file mode 100644
index 000000000000..334857e0832c
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_relation_references/result.json
@@ -0,0 +1,15 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 2,
+        "character": 4
+      },
+      "end": {
+        "line": 2,
+        "character": 8
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/model_unique_fields/result.json b/prisma-fmt/tests/references/scenarios/model_unique_fields/result.json
new file mode 100644
index 000000000000..6908574aac86
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_unique_fields/result.json
@@ -0,0 +1,15 @@
+[
+  {
+    "uri": "file:///path/to/schema.prisma",
+    "range": {
+      "start": {
+        "line": 11,
+        "character": 4
+      },
+      "end": {
+        "line": 11,
+        "character": 8
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/model_unique_fields/schema.prisma b/prisma-fmt/tests/references/scenarios/model_unique_fields/schema.prisma
new file mode 100644
index 000000000000..d0194d75b6e6
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/model_unique_fields/schema.prisma
@@ -0,0 +1,16 @@
+generator client {
+    provider        = "prisma-client-js"
+}
+
+datasource db {
+    provider = "postgres"
+    url      = env("DATABASE_URL")
+}
+
+model Compound {
+    id      String
+    name    String
+
+    @@unique([id, n<|>ame])
+}
+
diff --git a/prisma-fmt/tests/references/scenarios/view_as_type/a.prisma b/prisma-fmt/tests/references/scenarios/view_as_type/a.prisma
new file mode 100644
index 000000000000..ffd958d2799d
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_as_type/a.prisma
@@ -0,0 +1,16 @@
+generator client {
+  provider        = "prisma-client-js"
+  previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+  provider = "mongodb"
+  url      = env("DATABASE_URL")
+}
+
+view UserTwo {
+    id String @id @map("_id") @db.String
+
+    posts  Post   @relation(fields: [postId], references: [id])
+    postId String
+}
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/view_as_type/b.prisma b/prisma-fmt/tests/references/scenarios/view_as_type/b.prisma
new file mode 100644
index 000000000000..58f0d86dae09
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_as_type/b.prisma
@@ -0,0 +1,5 @@
+model Post {
+  id       String    @id @map("_id")
+  authorId String
+  UserTwo  User<|>Two[]
+}
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/view_as_type/result.json b/prisma-fmt/tests/references/scenarios/view_as_type/result.json
new file mode 100644
index 000000000000..d02299ab411e
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_as_type/result.json
@@ -0,0 +1,28 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 10,
+        "character": 5
+      },
+      "end": {
+        "line": 10,
+        "character": 12
+      }
+    }
+  },
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 3,
+        "character": 11
+      },
+      "end": {
+        "line": 3,
+        "character": 18
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/view_index_fields/result.json b/prisma-fmt/tests/references/scenarios/view_index_fields/result.json
new file mode 100644
index 000000000000..6908574aac86
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_index_fields/result.json
@@ -0,0 +1,15 @@
+[
+  {
+    "uri": "file:///path/to/schema.prisma",
+    "range": {
+      "start": {
+        "line": 11,
+        "character": 4
+      },
+      "end": {
+        "line": 11,
+        "character": 8
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/view_index_fields/schema.prisma b/prisma-fmt/tests/references/scenarios/view_index_fields/schema.prisma
new file mode 100644
index 000000000000..df22e6509fba
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_index_fields/schema.prisma
@@ -0,0 +1,16 @@
+generator client {
+    provider        = "prisma-client-js"
+}
+
+datasource db {
+    provider = "postgres"
+    url      = env("DATABASE_URL")
+}
+
+model Compound {
+    id      String
+    name    String
+
+    @@unique(fields: [id, n<|>ame])
+}
+
diff --git a/prisma-fmt/tests/references/scenarios/view_name/a.prisma b/prisma-fmt/tests/references/scenarios/view_name/a.prisma
new file mode 100644
index 000000000000..dfb42c85b4b8
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_name/a.prisma
@@ -0,0 +1,16 @@
+generator client {
+  provider        = "prisma-client-js"
+  previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+  provider = "mongodb"
+  url      = env("DATABASE_URL")
+}
+
+view User<|>Two {
+    id String @id @map("_id") @db.String
+
+    posts  Post   @relation(fields: [postId], references: [id])
+    postId String
+}
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/view_name/b.prisma b/prisma-fmt/tests/references/scenarios/view_name/b.prisma
new file mode 100644
index 000000000000..107a84502a96
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_name/b.prisma
@@ -0,0 +1,5 @@
+model Post {
+  id       String    @id @map("_id")
+  authorId String
+  UserTwo  UserTwo[]
+}
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/view_name/result.json b/prisma-fmt/tests/references/scenarios/view_name/result.json
new file mode 100644
index 000000000000..d02299ab411e
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_name/result.json
@@ -0,0 +1,28 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 10,
+        "character": 5
+      },
+      "end": {
+        "line": 10,
+        "character": 12
+      }
+    }
+  },
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 3,
+        "character": 11
+      },
+      "end": {
+        "line": 3,
+        "character": 18
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/view_relation_fields/a.prisma b/prisma-fmt/tests/references/scenarios/view_relation_fields/a.prisma
new file mode 100644
index 000000000000..8d0604d42728
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_relation_fields/a.prisma
@@ -0,0 +1,7 @@
+view Compound {
+    id      String
+    name    String
+    UserTwo UserTwo[]
+
+    @@id([id, name])
+}
diff --git a/prisma-fmt/tests/references/scenarios/view_relation_fields/b.prisma b/prisma-fmt/tests/references/scenarios/view_relation_fields/b.prisma
new file mode 100644
index 000000000000..5d7d155b70bb
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_relation_fields/b.prisma
@@ -0,0 +1,7 @@
+view UserTwo {
+    id String @id @map("_id")
+
+    compounds    Compound @relation(fields: [comp<|>oundId, compoundName], references: [id, name])
+    compoundId   String
+    compoundName String
+}
diff --git a/prisma-fmt/tests/references/scenarios/view_relation_fields/config.prisma b/prisma-fmt/tests/references/scenarios/view_relation_fields/config.prisma
new file mode 100644
index 000000000000..523204f6f2f5
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_relation_fields/config.prisma
@@ -0,0 +1,9 @@
+generator client {
+    provider        = "prisma-client-js"
+    previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+    provider = "postgres"
+    url      = env("DATABASE_URL")
+}
diff --git a/prisma-fmt/tests/references/scenarios/view_relation_fields/result.json b/prisma-fmt/tests/references/scenarios/view_relation_fields/result.json
new file mode 100644
index 000000000000..fbbe001172cf
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_relation_fields/result.json
@@ -0,0 +1,15 @@
+[
+  {
+    "uri": "file:///path/to/b.prisma",
+    "range": {
+      "start": {
+        "line": 4,
+        "character": 4
+      },
+      "end": {
+        "line": 4,
+        "character": 14
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/scenarios/view_relation_references/a.prisma b/prisma-fmt/tests/references/scenarios/view_relation_references/a.prisma
new file mode 100644
index 000000000000..8d0604d42728
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_relation_references/a.prisma
@@ -0,0 +1,7 @@
+view Compound {
+    id      String
+    name    String
+    UserTwo UserTwo[]
+
+    @@id([id, name])
+}
diff --git a/prisma-fmt/tests/references/scenarios/view_relation_references/b.prisma b/prisma-fmt/tests/references/scenarios/view_relation_references/b.prisma
new file mode 100644
index 000000000000..17e068b4d574
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_relation_references/b.prisma
@@ -0,0 +1,7 @@
+view UserTwo {
+    id String @id @map("_id")
+
+    compounds    Compound @relation(fields: [compoundId, compoundName], references: [id, n<|>ame])
+    compoundId   String
+    compoundName String
+}
diff --git a/prisma-fmt/tests/references/scenarios/view_relation_references/config.prisma b/prisma-fmt/tests/references/scenarios/view_relation_references/config.prisma
new file mode 100644
index 000000000000..523204f6f2f5
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_relation_references/config.prisma
@@ -0,0 +1,9 @@
+generator client {
+    provider        = "prisma-client-js"
+    previewFeatures = ["prismaSchemaFolder", "views"]
+}
+
+datasource db {
+    provider = "postgres"
+    url      = env("DATABASE_URL")
+}
diff --git a/prisma-fmt/tests/references/scenarios/view_relation_references/result.json b/prisma-fmt/tests/references/scenarios/view_relation_references/result.json
new file mode 100644
index 000000000000..334857e0832c
--- /dev/null
+++ b/prisma-fmt/tests/references/scenarios/view_relation_references/result.json
@@ -0,0 +1,15 @@
+[
+  {
+    "uri": "file:///path/to/a.prisma",
+    "range": {
+      "start": {
+        "line": 2,
+        "character": 4
+      },
+      "end": {
+        "line": 2,
+        "character": 8
+      }
+    }
+  }
+]
\ No newline at end of file
diff --git a/prisma-fmt/tests/references/test_api.rs b/prisma-fmt/tests/references/test_api.rs
new file mode 100644
index 000000000000..4fe7c0ffbc9d
--- /dev/null
+++ b/prisma-fmt/tests/references/test_api.rs
@@ -0,0 +1,149 @@
+use crate::helpers::load_schema_files;
+use once_cell::sync::Lazy;
+use std::{fmt::Write as _, io::Write as _};
+
+const CURSOR_MARKER: &str = "<|>";
+const SCENARIOS_PATH: &str = concat!(env!("CARGO_MANIFEST_DIR"), "/tests/references/scenarios");
+static UPDATE_EXPECT: Lazy<bool> = Lazy::new(|| std::env::var("UPDATE_EXPECT").is_ok());
+
+pub(crate) fn test_scenario(scenario_name: &str) {
+    let mut path = String::with_capacity(SCENARIOS_PATH.len() + 12);
+
+    let schema_files = {
+        write!(path, "{SCENARIOS_PATH}/{scenario_name}").unwrap();
+        load_schema_files(&path)
+    };
+
+    path.clear();
+    write!(path, "{SCENARIOS_PATH}/{scenario_name}/result.json").unwrap();
+    let expected_result = std::fs::read_to_string(&path).unwrap_or_else(|_| String::new());
+
+    let (initiating_file_uri, cursor_position, schema_files) = take_cursor(schema_files);
+
+    let params = lsp_types::ReferenceParams {
+        text_document_position: lsp_types::TextDocumentPositionParams {
+            text_document: lsp_types::TextDocumentIdentifier {
+                uri: initiating_file_uri.parse().unwrap(),
+            },
+            position: cursor_position,
+        },
+        work_done_progress_params: lsp_types::WorkDoneProgressParams { work_done_token: None },
+        partial_result_params: lsp_types::PartialResultParams {
+            partial_result_token: None,
+        },
+        context: lsp_types::ReferenceContext {
+            include_declaration: true,
+        },
+    };
+
+    let result = prisma_fmt::references(
+        serde_json::to_string_pretty(&schema_files).unwrap(),
+        &serde_json::to_string_pretty(&params).unwrap(),
+    );
+    // Prettify the JSON
+    let result =
+        serde_json::to_string_pretty(&serde_json::from_str::<Vec<lsp_types::Location>>(&result).unwrap()).unwrap();
+
+    if *UPDATE_EXPECT {
+        let mut file = std::fs::File::create(&path).unwrap(); // truncate
+        file.write_all(result.as_bytes()).unwrap();
+    } else if expected_result != result {
+        let chunks = dissimilar::diff(&expected_result, &result);
+        panic!(
+            r#"
+Snapshot comparison failed. Run the test again with UPDATE_EXPECT=1 in the environment to update the snapshot.
+
+===== EXPECTED ====
+{}
+====== FOUND ======
+{}
+======= DIFF ======
+{}
+"#,
+            expected_result,
+            result,
+            format_chunks(chunks),
+        );
+    }
+}
+
+fn format_chunks(chunks: Vec<dissimilar::Chunk>) -> String {
+    let mut buf = String::new();
+    for chunk in chunks {
+        let formatted = match chunk {
+            dissimilar::Chunk::Equal(text) => text.into(),
+            dissimilar::Chunk::Delete(text) => format!("\x1b[41m{text}\x1b[0m"),
+            dissimilar::Chunk::Insert(text) => format!("\x1b[42m{text}\x1b[0m"),
+        };
+        buf.push_str(&formatted);
+    }
+    buf
+}
+
+fn take_cursor(schema_files: Vec<(String, String)>) -> (String, lsp_types::Position, Vec<(String, String)>) {
+    let mut result = Vec::with_capacity(schema_files.len());
+    let mut file_and_pos = None;
+    for (file_name, content) in schema_files {
+        if let Some((pos, without_cursor)) = take_cursor_one(&content) {
+            file_and_pos = Some((file_name.clone(), pos));
+            result.push((file_name, without_cursor));
+        } else {
+            result.push((file_name, content));
+        }
+    }
+
+    let (file_name, position) = file_and_pos.expect("Could not find a cursor in any of the schema files");
+
+    (file_name, position, result)
+}
+
+fn take_cursor_one(schema: &str) -> Option<(lsp_types::Position, String)> {
+    let mut schema_without_cursor = String::with_capacity(schema.len() - 3);
+    let mut cursor_position = lsp_types::Position { character: 0, line: 0 };
+    let mut cursor_found = false;
+    for line in schema.lines() {
+        if !cursor_found {
+            if let Some(pos) = line.find(CURSOR_MARKER) {
+                cursor_position.character = pos as u32;
+                cursor_found = true;
+                schema_without_cursor.push_str(&line[..pos]);
+                schema_without_cursor.push_str(&line[pos + 3..]);
+                schema_without_cursor.push('\n');
+            } else {
+                schema_without_cursor.push_str(line);
+                schema_without_cursor.push('\n');
+                cursor_position.line += 1;
+            }
+        } else {
+            schema_without_cursor.push_str(line);
+            schema_without_cursor.push('\n');
+        }
+    }
+
+    if !cursor_found {
+        return None;
+    }
+    // remove extra newline
+    schema_without_cursor.truncate(schema_without_cursor.len() - 1);
+
+    Some((cursor_position, schema_without_cursor))
+}
+
+#[test]
+fn take_cursor_works() {
+    let schema = r#"
+        model Test {
+            id Int @id @map(<|>)
+        }
+    "#;
+    let expected_schema = r#"
+        model Test {
+            id Int @id @map()
+        }
+    "#;
+
+    let (pos, schema) = take_cursor_one(schema).unwrap();
+    assert_eq!(pos.line, 2);
+    assert_eq!(pos.character, 28);
+    assert_eq!(schema, expected_schema);
+}
diff --git a/prisma-fmt/tests/references/tests.rs b/prisma-fmt/tests/references/tests.rs
new file mode 100644
index 000000000000..c32a833f12be
--- /dev/null
+++ b/prisma-fmt/tests/references/tests.rs
@@ -0,0 +1,31 @@
+use super::test_api::test_scenario;
+
+macro_rules! scenarios {
+    ($($scenario_name:ident)+) => {
+        $(
+            #[test]
+            fn $scenario_name() {
+                test_scenario(stringify!($scenario_name))
+            }
+        )*
+    }
+}
+
+scenarios! {
+    composite_type_as_type
+    composite_type_name
+    enum_as_type
+    enum_name
+    model_as_type
+    model_name
+    model_relation_fields
+    model_relation_references
+    model_unique_fields
+    view_as_type
+    view_index_fields
+    view_name
+    view_relation_fields
+    view_relation_references
+    datasource_as_attribute
+    datasource_name
+}
diff --git a/prisma-fmt/tests/references_tests.rs b/prisma-fmt/tests/references_tests.rs
new file mode 100644
index 000000000000..7b6416981551
--- /dev/null
+++ b/prisma-fmt/tests/references_tests.rs
@@ -0,0 +1,2 @@
+mod helpers;
+mod references;
diff --git a/prisma-fmt/tests/text_document_completion/scenarios/default_map_mssql_in_arg_value/result.json b/prisma-fmt/tests/text_document_completion/scenarios/default_map_mssql_in_arg_value/result.json
new file mode 100644
index 000000000000..cd41656e5303
--- /dev/null
+++ b/prisma-fmt/tests/text_document_completion/scenarios/default_map_mssql_in_arg_value/result.json
@@ -0,0 +1,4 @@
+{
+  "isIncomplete": false,
+  "items": []
+}
\ No newline at end of file
diff --git a/prisma-fmt/tests/text_document_completion/scenarios/default_map_mssql_in_arg_value/schema.prisma b/prisma-fmt/tests/text_document_completion/scenarios/default_map_mssql_in_arg_value/schema.prisma
new file mode 100644
index 000000000000..444e86c947ab
--- /dev/null
+++ b/prisma-fmt/tests/text_document_completion/scenarios/default_map_mssql_in_arg_value/schema.prisma
@@ -0,0 +1,10 @@
+datasource db {
+    provider = "sqlserver"
+    url      = env("DATABASE_URL")
+}
+
+model User {
+    id String @id
+
+    postId String @unique @default("defaul<|>t_value")
+}
diff --git a/prisma-fmt/tests/text_document_completion/scenarios/referential_actions_in_progress/result.json b/prisma-fmt/tests/text_document_completion/scenarios/referential_actions_in_progress/result.json
index 0d1b8ac130e1..8b1a51e6b215 100644
--- a/prisma-fmt/tests/text_document_completion/scenarios/referential_actions_in_progress/result.json
+++ b/prisma-fmt/tests/text_document_completion/scenarios/referential_actions_in_progress/result.json
@@ -5,21 +5,6 @@
       "label": "Cascade",
       "kind": 13,
       "detail": "Delete the child records when the parent record is deleted."
-    },
-    {
-      "label": "NoAction",
-      "kind": 13,
-      "detail": "Prevent deleting a parent record as long as it is referenced."
-    },
-    {
-      "label": "SetNull",
-      "kind": 13,
-      "detail": "Set the referencing fields to NULL when the referenced record is deleted."
-    },
-    {
-      "label": "SetDefault",
-      "kind": 13,
-      "detail": "Set the referencing field's value to the default when the referenced record is deleted."
     }
   ]
 }
\ No newline at end of file
diff --git a/prisma-fmt/tests/text_document_completion/scenarios/referential_actions_in_progress_2/result.json b/prisma-fmt/tests/text_document_completion/scenarios/referential_actions_in_progress_2/result.json
index 24180484bbd2..5609cdf1e9eb 100644
--- a/prisma-fmt/tests/text_document_completion/scenarios/referential_actions_in_progress_2/result.json
+++ b/prisma-fmt/tests/text_document_completion/scenarios/referential_actions_in_progress_2/result.json
@@ -1,30 +1,10 @@
 {
   "isIncomplete": false,
   "items": [
-    {
-      "label": "Cascade",
-      "kind": 13,
-      "detail": "Delete the child records when the parent record is deleted."
-    },
     {
       "label": "Restrict",
       "kind": 13,
       "detail": "Prevent deleting a parent record as long as it is referenced."
-    },
-    {
-      "label": "NoAction",
-      "kind": 13,
-      "detail": "Prevent deleting a parent record as long as it is referenced."
-    },
-    {
-      "label": "SetNull",
-      "kind": 13,
-      "detail": "Set the referencing fields to NULL when the referenced record is deleted."
-    },
-    {
-      "label": "SetDefault",
-      "kind": 13,
-      "detail": "Set the referencing field's value to the default when the referenced record is deleted."
     }
   ]
 }
\ No newline at end of file
diff --git a/prisma-fmt/tests/text_document_completion/tests.rs b/prisma-fmt/tests/text_document_completion/tests.rs
index c1332e6f9d41..1066c32fdfbb 100644
--- a/prisma-fmt/tests/text_document_completion/tests.rs
+++ b/prisma-fmt/tests/text_document_completion/tests.rs
@@ -15,6 +15,7 @@ scenarios! {
     argument_after_trailing_comma
     default_map_end_of_args_list
     default_map_mssql
+    default_map_mssql_in_arg_value
     default_map_mssql_multifile
     empty_schema
     extended_indexes_basic
diff --git a/prisma-schema-wasm/src/lib.rs b/prisma-schema-wasm/src/lib.rs
index c331170b2366..5cbb08067877 100644
--- a/prisma-schema-wasm/src/lib.rs
+++ b/prisma-schema-wasm/src/lib.rs
@@ -86,7 +86,7 @@ pub fn preview_features() -> String {
 }
 
 /// The API is modelled on an LSP [completion
-/// request](https://github.com/microsoft/language-server-protocol/blob/gh-pages/_specifications/specification-3-16.md#textDocument_completion).
+/// request](https://github.com/microsoft/language-server-protocol/blob/gh-pages/_specifications/specification-3-16.md#completion-request-leftwards_arrow_with_hook).
 /// Input and output are both JSON, the request being a `CompletionParams` object and the response
 /// being a `CompletionList` object.
 #[wasm_bindgen]
@@ -96,7 +96,7 @@ pub fn text_document_completion(schema_files: String, params: String) -> String
 }
 
 /// This API is modelled on an LSP [code action
-/// request](https://github.com/microsoft/language-server-protocol/blob/gh-pages/_specifications/specification-3-16.md#textDocument_codeAction=).
+/// request](https://github.com/microsoft/language-server-protocol/blob/gh-pages/_specifications/specification-3-16.md#code-action-request-leftwards_arrow_with_hook).
 /// Input and output are both JSON, the request being a
 /// `CodeActionParams` object and the response being a list of
 /// `CodeActionOrCommand` objects.
@@ -106,6 +106,17 @@ pub fn code_actions(schema: String, params: String) -> String {
     prisma_fmt::code_actions(schema, &params)
 }
 
+/// This API is modelled on an LSP [references
+/// request](https://github.com/microsoft/language-server-protocol/blob/gh-pages/_specifications/specification-3-16.md#find-references-request-leftwards_arrow_with_hook).
+/// Input and output are both JSON, the request being a
+/// `CodeActionParams` object and the response being a list of
+/// `CodeActionOrCommand` objects.
+#[wasm_bindgen]
+pub fn references(schema: String, params: String) -> String {
+    register_panic_hook();
+    prisma_fmt::references(schema, &params)
+}
+
 /// Trigger a panic inside the wasm module. This is only useful in development for testing panic
 /// handling.
 #[wasm_bindgen]
diff --git a/psl/parser-database/src/walkers.rs b/psl/parser-database/src/walkers.rs
index ee1058f9c2fc..dd8cf2504eae 100644
--- a/psl/parser-database/src/walkers.rs
+++ b/psl/parser-database/src/walkers.rs
@@ -75,7 +75,15 @@ impl crate::ParserDatabase {
             .flat_map(move |(file_id, _, _, ast)| ast.iter_tops().map(move |(top_id, top)| (file_id, top_id, top)))
     }
 
-    /// Intern any top by name.
+    /// Find the datasource by name.
+    pub fn find_source(&self, name: &str) -> Option<(FileId, ast::TopId)> {
+        self.interner
+            .lookup(name)
+            .and_then(|name_id| self.names.datasources.get(&name_id))
+            .map(|(file_id, top_id)| (*file_id, *top_id))
+    }
+
+    /// Find any top by name.
     pub fn find_top<'db>(&'db self, name: &str) -> Option<TopWalker<'db>> {
         self.interner
             .lookup(name)
diff --git a/psl/psl-core/src/builtin_connectors/mssql_datamodel_connector.rs b/psl/psl-core/src/builtin_connectors/mssql_datamodel_connector.rs
index 6eb554787b84..d55b1429e23a 100644
--- a/psl/psl-core/src/builtin_connectors/mssql_datamodel_connector.rs
+++ b/psl/psl-core/src/builtin_connectors/mssql_datamodel_connector.rs
@@ -262,7 +262,10 @@ impl Connector for MsSqlDatamodelConnector {
     ) {
         if let ast::SchemaPosition::Model(
             _model_id,
-            ast::ModelPosition::Field(_, ast::FieldPosition::Attribute("default", _, None)),
+            ast::ModelPosition::Field(
+                _,
+                ast::FieldPosition::Attribute("default", _, ast::AttributePosition::Attribute),
+            ),
         ) = position
         {
             completions.items.push(CompletionItem {
diff --git a/psl/psl-core/src/builtin_connectors/postgres_datamodel_connector.rs b/psl/psl-core/src/builtin_connectors/postgres_datamodel_connector.rs
index 22badc3e0363..fa1559c33c2b 100644
--- a/psl/psl-core/src/builtin_connectors/postgres_datamodel_connector.rs
+++ b/psl/psl-core/src/builtin_connectors/postgres_datamodel_connector.rs
@@ -495,7 +495,7 @@ impl Connector for PostgresDatamodelConnector {
                 ast::ModelPosition::ModelAttribute(
                     "index",
                     attr_id,
-                    ast::AttributePosition::FunctionArgument(field_name, "ops"),
+                    ast::AttributePosition::FunctionArgument(field_name, "ops", _),
                 ),
             ) => {
                 // let's not care about composite field indices yet
diff --git a/psl/schema-ast/src/ast/argument.rs b/psl/schema-ast/src/ast/argument.rs
index 9272196780e9..a96f05cfa384 100644
--- a/psl/schema-ast/src/ast/argument.rs
+++ b/psl/schema-ast/src/ast/argument.rs
@@ -68,6 +68,13 @@ impl Argument {
     pub fn is_unnamed(&self) -> bool {
         self.name.is_none()
     }
+
+    pub fn name(&self) -> Option<&str> {
+        match &self.name {
+            Some(ident) => Some(ident.name.as_str()),
+            None => None,
+        }
+    }
 }
 
 impl WithSpan for Argument {
diff --git a/psl/schema-ast/src/ast/find_at_position/attribute.rs b/psl/schema-ast/src/ast/find_at_position/attribute.rs
index 07dea044176c..fd1ea3ffb31d 100644
--- a/psl/schema-ast/src/ast/find_at_position/attribute.rs
+++ b/psl/schema-ast/src/ast/find_at_position/attribute.rs
@@ -9,50 +9,60 @@ pub enum AttributePosition<'ast> {
     Attribute,
     /// In an argument. (argument name)
     Argument(&'ast str),
-    /// In an function argument. (function name, argument name)
-    FunctionArgument(&'ast str, &'ast str),
+    /// In an argument's value. (argument name, value)
+    ArgumentValue(Option<&'ast str>, String),
+    /// In an function argument. (function name, argument name, argument value)
+    FunctionArgument(&'ast str, &'ast str, String),
 }
 
 impl<'ast> AttributePosition<'ast> {
     pub(crate) fn new(attr: &'ast ast::Attribute, position: usize) -> Self {
-        if attr.span().contains(position) {
-            // We can't go by Span::contains() because we also care about the empty space
-            // between arguments and that's hard to capture in the pest grammar.
-            let mut spans: Vec<(Option<&str>, ast::Span)> = attr
-                .arguments
-                .iter()
-                .map(|arg| (arg.name.as_ref().map(|n| n.name.as_str()), arg.span()))
-                .chain(
-                    attr.arguments
-                        .empty_arguments
-                        .iter()
-                        .map(|arg| (Some(arg.name.name.as_str()), arg.name.span())),
-                )
-                .collect();
-
-            spans.sort_by_key(|(_, span)| span.start);
-
-            let mut arg_name = None;
-            for (name, _) in spans.iter().take_while(|(_, span)| span.start < position) {
-                arg_name = Some(*name);
-            }
+        let mut spans: Vec<(Option<&str>, ast::Span)> = attr
+            .arguments
+            .iter()
+            .map(|arg| (arg.name.as_ref().map(|n| n.name.as_str()), arg.span()))
+            .chain(
+                attr.arguments
+                    .empty_arguments
+                    .iter()
+                    .map(|arg| (Some(arg.name.name.as_str()), arg.name.span())),
+            )
+            .collect();
 
-            // If the cursor is after a trailing comma, we're not in an argument.
-            if let Some(span) = attr.arguments.trailing_comma {
-                if position > span.start {
-                    arg_name = None;
-                }
+        spans.sort_by_key(|(_, span)| span.start);
+
+        let mut arg_name = None;
+        for (name, _) in spans.iter().take_while(|(_, span)| span.start < position) {
+            arg_name = Some(*name);
+        }
+
+        // If the cursor is after a trailing comma, we're not in an argument.
+        if let Some(span) = attr.arguments.trailing_comma {
+            if position > span.start {
+                arg_name = None;
             }
+        }
 
-            if let Some(arg_name) = arg_name.flatten() {
-                return Self::Argument(arg_name);
+        if let Some(arg) = attr.arguments.iter().find(|arg| arg.span().contains(position)) {
+            if let ExpressionPosition::FunctionArgument(fun, name) = ExpressionPosition::new(&arg.value, position) {
+                return Self::FunctionArgument(fun, name, arg.value.to_string());
             }
 
-            if let Some(arg) = attr.arguments.iter().find(|arg| arg.span().contains(position)) {
-                if let ExpressionPosition::FunctionArgument(fun, name) = ExpressionPosition::new(&arg.value, position) {
-                    return Self::FunctionArgument(fun, name);
+            if arg.value.is_array() {
+                let arr = arg.value.as_array().unwrap();
+                let expr = arr.0.iter().find(|expr| expr.span().contains(position));
+                if let Some(expr) = expr {
+                    return Self::ArgumentValue(arg.name(), expr.to_string());
                 }
             }
+
+            return Self::ArgumentValue(arg.name(), arg.value.to_string());
+
+            // Self::ArgumentValue(arg_name, ())
+        }
+
+        if let Some(arg_name) = arg_name.flatten() {
+            return Self::Argument(arg_name);
         }
 
         Self::Attribute
diff --git a/psl/schema-ast/src/ast/find_at_position/field.rs b/psl/schema-ast/src/ast/find_at_position/field.rs
index 3c358f3c6da7..97995fad1678 100644
--- a/psl/schema-ast/src/ast/find_at_position/field.rs
+++ b/psl/schema-ast/src/ast/find_at_position/field.rs
@@ -1,6 +1,6 @@
 use crate::ast::{self};
 
-use super::{WithName, WithSpan};
+use super::{AttributePosition, WithName, WithSpan};
 
 /// In a scalar field.
 #[derive(Debug)]
@@ -33,7 +33,8 @@ pub enum FieldPosition<'ast> {
     ///     field Float
     /// }
     /// ```
-    Attribute(&'ast str, usize, Option<&'ast str>),
+    // Attribute(&'ast str, usize, Option<&'ast str>),
+    Attribute(&'ast str, usize, AttributePosition<'ast>),
 }
 
 impl<'ast> FieldPosition<'ast> {
@@ -50,32 +51,7 @@ impl<'ast> FieldPosition<'ast> {
             if attr.span().contains(position) {
                 // We can't go by Span::contains() because we also care about the empty space
                 // between arguments and that's hard to capture in the pest grammar.
-                let mut spans: Vec<(Option<&str>, ast::Span)> = attr
-                    .arguments
-                    .iter()
-                    .map(|arg| (arg.name.as_ref().map(|n| n.name.as_str()), arg.span()))
-                    .chain(
-                        attr.arguments
-                            .empty_arguments
-                            .iter()
-                            .map(|arg| (Some(arg.name.name.as_str()), arg.name.span())),
-                    )
-                    .collect();
-                spans.sort_by_key(|(_, span)| span.start);
-                let mut arg_name = None;
-
-                for (name, _) in spans.iter().take_while(|(_, span)| span.start < position) {
-                    arg_name = Some(*name);
-                }
-
-                // If the cursor is after a trailing comma, we're not in an argument.
-                if let Some(span) = attr.arguments.trailing_comma {
-                    if position > span.start {
-                        arg_name = None;
-                    }
-                }
-
-                return FieldPosition::Attribute(attr.name(), attr_idx, arg_name.flatten());
+                return FieldPosition::Attribute(attr.name(), attr_idx, AttributePosition::new(attr, position));
             }
         }