simstudioai · icecrasher321 · Sep 17, 2026 · Sep 16, 2026 · Sep 16, 2026 · Sep 16, 2026
diff --git a/apps/sim/lib/file-parsers/index.ts b/apps/sim/lib/file-parsers/index.ts
@@ -176,7 +176,7 @@ export async function parseBuffer(
     }
 
     const kind = sniffFileKind(buffer, normalizedExtension)
-    const route = reconcileParserRoute(normalizedExtension, kind)
+    const route = reconcileParserRoute(normalizedExtension, kind, options)
     const parser = PARSERS.get(route.extension)
 
     if (!parser?.parseBuffer) {

diff --git a/apps/sim/lib/file-parsers/sniff.test.ts b/apps/sim/lib/file-parsers/sniff.test.ts
@@ -308,6 +308,38 @@ describe('parseBuffer reconciles the extension with the sniffed bytes', () => {
     expect(result.metadata?.detectedType).toBe('html')
   })
 
+  it.each([
+    '<!doctype html><html><head><script src="./app.js"></script></head><body><div id="root"></div></body></html>',
+    '{\\rtf1\\ansi Source-format example}',
+  ])(
+    'preserves textual markup when the caller supplies a canonical text artifact',
+    async (content) => {
+      const result = await parseBuffer(Buffer.from(content), 'txt', { textMode: 'literal' })
+
+      expect(result.content).toBe(content)
+      expect(result.metadata?.detectedType).toBeUndefined()
+    }
+  )
+
+  it('keeps literal-text handling scoped to txt artifacts', async () => {
+    const result = await parseBuffer(
+      Buffer.from('<!doctype html><html><body><p>Readable page</p></body></html>'),
+      'html',
+      { textMode: 'literal' }
+    )
+
+    expect(result.content).toContain('Readable page')
+    expect(result.content).not.toContain('<html>')
+    await expect(
+      parseBuffer(Buffer.from('<!doctype html><html><body>403 Forbidden</body></html>'), 'json', {
+        textMode: 'literal',
+      })
+    ).rejects.toMatchObject({ code: 'invalid_format' })
+    await expect(parseBuffer(oleBinary(), 'txt', { textMode: 'literal' })).rejects.toMatchObject({
+      code: 'invalid_format',
+    })
+  })
+
   it('extracts a docx labelled .xlsx through the Word parser', async () => {
     const result = await parseBuffer(await buildDocx('Office Relocation'), 'xlsx')
 

diff --git a/apps/sim/lib/file-parsers/sniff.ts b/apps/sim/lib/file-parsers/sniff.ts
@@ -1,5 +1,6 @@
 import { FileParserError } from '@/lib/file-parsers/errors'
 import { isEncryptedOoxmlContainer } from '@/lib/file-parsers/ooxml-encryption'
+import type { FileParseOptions } from '@/lib/file-parsers/types'
 import { decodeTextBuffer, detectBomlessUtf16 } from '@/lib/file-parsers/utils'
 import { isZipShaped } from '@/lib/file-parsers/zip-guard'
 
@@ -307,7 +308,18 @@ function invalidFormat(extension: string, kind: SniffedKind): FileParserError {
  * text (as CSV under a spreadsheet extension), and an OLE2 file under a modern
  * Word extension is the legacy `.doc` parser's job. Legacy `.ppt` has no reader.
  */
-export function reconcileParserRoute(extension: string, kind: SniffedKind): ParserRoute {
+export function reconcileParserRoute(
+  extension: string,
+  kind: SniffedKind,
+  options: Pick<FileParseOptions, 'textMode'> = {}
+): ParserRoute {
+  if (
+    extension === 'txt' &&
+    options.textMode === 'literal' &&
+    (kind === 'html' || kind === 'rtf')
+  ) {
+    return { extension }
+  }
   if (kind === 'rtf') {
     throw new FileParserError(
       'unsupported_type',

diff --git a/apps/sim/lib/file-parsers/types.ts b/apps/sim/lib/file-parsers/types.ts
@@ -33,6 +33,8 @@ export interface FileParseResult {
 
 export interface FileParseOptions {
   signal?: AbortSignal
+  /** Preserve textual markup in a canonical .txt artifact instead of interpreting it as HTML or RTF. */
+  textMode?: 'literal'
   /** Complete PDF extraction rejects safety limits instead of returning preview text. */
   pdfTextMode?: 'preview' | 'complete'
 }