diff --git a/src/utils/eventsource-parser.mjs b/src/utils/eventsource-parser.mjs
index 802a17f0..436700ea 100644
--- a/src/utils/eventsource-parser.mjs
+++ b/src/utils/eventsource-parser.mjs
@@ -10,6 +10,7 @@ function createParser(onParse) {
   let eventName
   let data
   let extra
+  let discardTrailingNewline
   reset()
   return {
     feed,
@@ -24,6 +25,7 @@ function createParser(onParse) {
     eventId = void 0
     eventName = void 0
     data = ''
+    discardTrailingNewline = false
   }
 
   function feed(chunk) {
@@ -35,7 +37,6 @@ function createParser(onParse) {
     isFirstChunk = false
     const length = buffer.length
     let position = 0
-    let discardTrailingNewline = false
     while (position < length) {
       if (discardTrailingNewline) {
         if (buffer[position] === '\n') {
@@ -46,7 +47,11 @@ function createParser(onParse) {
       let lineLength = -1
       let fieldLength = startingFieldLength
       let character
-      for (let index = startingPosition; lineLength < 0 && index < length; ++index) {
+      for (
+        let index = position + startingPosition;
+        lineLength < 0 && index < length;
+        ++index
+      ) {
         character = buffer[index]
         if (character === ':' && fieldLength < 0) {
           fieldLength = index - position
diff --git a/tests/unit/utils/eventsource-parser.test.mjs b/tests/unit/utils/eventsource-parser.test.mjs
index c999454b..debe86ee 100644
--- a/tests/unit/utils/eventsource-parser.test.mjs
+++ b/tests/unit/utils/eventsource-parser.test.mjs
@@ -6,6 +6,13 @@ const encoder = new TextEncoder()
 
 const toBytes = (text) => encoder.encode(text)
 
+const parseChunks = (...chunks) => {
+  const parsed = []
+  const parser = createParser((event) => parsed.push(event))
+  for (const chunk of chunks) parser.feed(chunk)
+  return parsed
+}
+
 test('createParser parses basic SSE event data', () => {
   const parsed = []
   const parser = createParser((event) => parsed.push(event))
@@ -72,6 +79,50 @@ test('createParser handles \\r\\n line endings', () => {
   assert.equal(parsed[0].data, 'hello\nworld')
 })
 
+test('createParser keeps CRLF separators intact across chunks', () => {
+  const parsed = []
+  const parser = createParser((event) => parsed.push(event))
+
+  parser.feed(toBytes('data: hello\r'))
+  parser.feed(toBytes('\ndata: world\r'))
+  parser.feed(toBytes('\n\r'))
+  parser.feed(toBytes('\n'))
+
+  assert.equal(parsed.length, 1)
+  assert.equal(parsed[0].data, 'hello\nworld')
+})
+
+test('createParser preserves an LF delimiter after a complete CRLF line', () => {
+  const parsed = parseChunks(toBytes('data: a\r\n'), toBytes('\n'))
+
+  assert.equal(parsed.length, 1)
+  assert.equal(parsed[0].data, 'a')
+})
+
+test('createParser preserves mixed CRLF and LF event lines across chunks', () => {
+  const parsed = parseChunks(toBytes('data: a\r\ndata: b\n'), toBytes('\n'))
+
+  assert.equal(parsed.length, 1)
+  assert.equal(parsed[0].data, 'a\nb')
+})
+
+test('createParser preserves split CRLF state across empty chunks', () => {
+  const parsed = parseChunks(toBytes('data: a\r'), toBytes(''), toBytes('\n\n'))
+
+  assert.equal(parsed.length, 1)
+  assert.equal(parsed[0].data, 'a')
+})
+
+test('createParser produces the same events at every single chunk boundary', () => {
+  const stream = toBytes('data: alpha\r\n\ndata: beta\ndata: gamma\r\r')
+  const expected = parseChunks(stream)
+
+  for (let split = 0; split <= stream.length; ++split) {
+    const actual = parseChunks(stream.slice(0, split), stream.slice(split))
+    assert.deepEqual(actual, expected, `split at byte ${split}`)
+  }
+})
+
 test('createParser handles \\r only line endings', () => {
   const parsed = []
   const parser = createParser((event) => parsed.push(event))