diff --git a/index/shard_builder.go b/index/shard_builder.go
index 3f76cc8e..8a3057b6 100644
--- a/index/shard_builder.go
+++ b/index/shard_builder.go
@@ -37,11 +37,6 @@ var _ = log.Println
 
 const ngramSize = 3
 
-// MaxFileSize 1 MB; match https://sourcegraph.sgdev.org/github.com/sourcegraph/sourcegraph/-/blob/cmd/symbols/internal/symbols/search.go#L22
-// NOTE: if you change this, you must also update gitIndex to use the same value when fetching the repo.
-// Change here as well, if you're changing the value https://sourcegraph.com/github.com/sourcegraph/zoekt/-/blob/cmd/zoekt-sourcegraph-indexserver/main.go?L167-169
-const MaxFileSize = 1 << 20
-
 type searchableString struct {
 	data []byte
 }
@@ -401,7 +396,16 @@ func (b *ShardBuilder) addSymbols(symbols []*zoekt.Symbol) {
 }
 
 func DetermineLanguageIfUnknown(doc *Document) {
-	if doc.Language == "" {
+	if doc.Language != "" {
+		return
+	}
+
+	if doc.SkipReason != "" {
+		// If this document has been skipped, it's likely very large, or it's a non-code file like binary.
+		// In this case, we just guess the language based on file name to avoid examining the contents.
+		// Note: passing nil content is allowed by the go-enry contract (the underlying library we use here).
+		doc.Language = languages.GetLanguage(doc.Name, nil)
+	} else {
 		doc.Language = languages.GetLanguage(doc.Name, doc.Content)
 	}
 }
@@ -410,9 +414,7 @@ func DetermineLanguageIfUnknown(doc *Document) {
 func (b *ShardBuilder) Add(doc Document) error {
 	hasher := crc64.New(crc64.MakeTable(crc64.ISO))
 
-	if len(doc.Content) > MaxFileSize {
-		doc.SkipReason = fmt.Sprintf("file size %d exceeds maximum size %d", len(doc.Content), MaxFileSize)
-	} else if idx := bytes.IndexByte(doc.Content, 0); idx >= 0 {
+	if idx := bytes.IndexByte(doc.Content, 0); idx >= 0 {
 		doc.SkipReason = fmt.Sprintf("binary content at byte offset %d", idx)
 	}
 
diff --git a/index/shard_builder_test.go b/index/shard_builder_test.go
index e66b3cf8..662ac8cd 100644
--- a/index/shard_builder_test.go
+++ b/index/shard_builder_test.go
@@ -47,3 +47,49 @@ func TestShardName(t *testing.T) {
 		})
 	}
 }
+
+func TestDetermineLanguageIfUnknown(t *testing.T) {
+	tests := []struct {
+		name        string
+		doc         Document
+		wantLang    string
+		skipContent bool
+	}{
+		{
+			name: "already has language",
+			doc: Document{
+				Name:     "test.java",
+				Language: "Go",
+				Content:  []byte("package main"),
+			},
+			wantLang: "Go",
+		},
+		{
+			name: "skipped file",
+			doc: Document{
+				Name:       "large.js",
+				SkipReason: "too large",
+				Content:    []byte(notIndexedMarker + "too large"),
+			},
+			wantLang: "JavaScript",
+		},
+		{
+			name: "skipped file with unknown extension",
+			doc: Document{
+				Name:       "deadb33f",
+				SkipReason: "binary",
+				Content:    []byte(notIndexedMarker + "binary"),
+			},
+			wantLang: "",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			DetermineLanguageIfUnknown(&tt.doc)
+			if tt.doc.Language != tt.wantLang {
+				t.Errorf("DetermineLanguageIfUnknown() got language = %v, want %v", tt.doc.Language, tt.wantLang)
+			}
+		})
+	}
+}