fix(textract): forward URL documents and fix ambiguous error status

- textract_analyze_expense/textract_analyze_id now fall back to filePath/filePathBack when the document input is a URL string rather than an uploaded file object, so advanced "File reference" URL inputs actually reach the API (Cursor Bugbot)
- mapTextractSdkError defaults to 500 (not 400) when the AWS SDK error has no HTTP status, since that implies a server-side/network failure rather than a bad request (Greptile)

Author: waleedlatif1

apps/sim/app/api/tools/textract/shared.test.ts

@@ -78,9 +78,9 @@ describe('mapTextractSdkError', () => {
     expect(mapped.status).toBe(500)
   })
 
-  it('defaults to 400 when the SDK gives no http status', () => {
+  it('defaults to 500 when the SDK gives no http status, since that implies a server-side failure', () => {
     const mapped = mapTextractSdkError({ message: 'Unknown failure' }, false)
-    expect(mapped.status).toBe(400)
+    expect(mapped.status).toBe(500)
   })
 })
 

apps/sim/app/api/tools/textract/shared.ts

@@ -75,7 +75,9 @@ export function mapTextractSdkError(
     return new TextractRouteError(`This document format is not supported.${hint}`, 400)
   }
 
-  const status = err.$metadata?.httpStatusCode || 400
+  // No status at all (e.g. a network failure before AWS responded) is a server-side problem, not a
+  // bad request — default to 500 so tool-execution retry logic still treats it as retryable.
+  const status = err.$metadata?.httpStatusCode ?? 500
   return new TextractRouteError(err.message || 'Textract API error', status)
 }
 

apps/sim/blocks/blocks/textract.ts

@@ -199,6 +199,18 @@ export const TextractBlock: BlockConfig<TextractParserOutput> = {
   },
 }
 
+/**
+ * Resolves a canonical document input to either an uploaded file object or a plain URL string.
+ * `normalizeFileInput` only recognizes file objects (or JSON-serialized file references) — a raw
+ * URL typed into the "File reference" advanced field falls through to `filePath` instead.
+ */
+function resolveDocumentParam(value: unknown): { file?: object; filePath?: string } {
+  const file = normalizeFileInput(value, { single: true })
+  if (file) return { file }
+  if (typeof value === 'string' && value.trim() !== '') return { filePath: value.trim() }
+  return {}
+}
+
 function requireAwsCredentials(params: Record<string, unknown>) {
   const accessKeyId = typeof params.accessKeyId === 'string' ? params.accessKeyId.trim() : ''
   const secretAccessKey =
@@ -350,17 +362,18 @@ export const TextractV2Block: BlockConfig<TextractParserOutput> = {
         const operation = params.operation || 'analyze_document'
 
         if (operation === 'analyze_id') {
-          const file = normalizeFileInput(params.document, { single: true })
-          if (!file) throw new Error('Identity document is required')
+          const front = resolveDocumentParam(params.document)
+          if (!front.file && !front.filePath) throw new Error('Identity document is required')
 
           const parameters: Record<string, unknown> = {
             accessKeyId,
             secretAccessKey,
             region,
-            file,
+            ...front,
           }
-          const fileBack = normalizeFileInput(params.documentBack, { single: true })
-          if (fileBack) parameters.fileBack = fileBack
+          const back = resolveDocumentParam(params.documentBack)
+          if (back.file) parameters.fileBack = back.file
+          else if (back.filePath) parameters.filePathBack = back.filePath
           return parameters
         }
 
@@ -377,7 +390,13 @@ export const TextractV2Block: BlockConfig<TextractParserOutput> = {
             throw new Error('S3 URI is required for multi-page processing')
           }
           parameters.s3Uri = params.s3Uri.trim()
+        } else if (operation === 'analyze_expense') {
+          // textract_analyze_expense accepts filePath as a fallback for URL-based documents.
+          const resolved = resolveDocumentParam(params.document)
+          if (!resolved.file && !resolved.filePath) throw new Error('Document file is required')
+          Object.assign(parameters, resolved)
         } else {
+          // textract_parser_v2 (analyze_document) has no filePath param — file only.
           const file = normalizeFileInput(params.document, { single: true })
           if (!file) throw new Error('Document file is required')
           parameters.file = file

apps/sim/tools/textract/analyze-expense.ts

@@ -99,6 +99,12 @@ export const textractAnalyzeExpenseTool: ToolConfig<
       visibility: 'hidden',
       description: 'Invoice or receipt to be processed (JPEG, PNG, or single-page PDF).',
     },
+    filePath: {
+      type: 'string',
+      required: false,
+      visibility: 'hidden',
+      description: 'URL to an invoice or receipt to be processed, if not uploaded directly.',
+    },
     s3Uri: {
       type: 'string',
       required: false,
@@ -125,11 +131,12 @@ export const textractAnalyzeExpenseTool: ToolConfig<
 
       if (processingMode === 'async') {
         requestBody.s3Uri = params.s3Uri?.trim()
-      } else {
-        if (!params.file || typeof params.file !== 'object') {
-          throw new Error('Document file is required for single-page processing')
-        }
+      } else if (params.file && typeof params.file === 'object') {
         requestBody.file = params.file
+      } else if (params.filePath && params.filePath.trim() !== '') {
+        requestBody.filePath = params.filePath.trim()
+      } else {
+        throw new Error('Document is required for single-page processing')
       }
 
       return requestBody

apps/sim/tools/textract/analyze-id.ts

@@ -37,12 +37,24 @@ export const textractAnalyzeIdTool: ToolConfig<TextractAnalyzeIdV2Input, Textrac
         visibility: 'hidden',
         description: 'Front of the identity document (JPEG, PNG, or PDF).',
       },
+      filePath: {
+        type: 'string',
+        required: false,
+        visibility: 'hidden',
+        description: 'URL to the front of the identity document, if not uploaded directly.',
+      },
       fileBack: {
         type: 'file',
         required: false,
         visibility: 'hidden',
         description: 'Back of the identity document, if applicable (JPEG, PNG, or PDF).',
       },
+      filePathBack: {
+        type: 'string',
+        required: false,
+        visibility: 'hidden',
+        description: 'URL to the back of the identity document, if not uploaded directly.',
+      },
     },
 
     request: {
@@ -53,19 +65,24 @@ export const textractAnalyzeIdTool: ToolConfig<TextractAnalyzeIdV2Input, Textrac
         Accept: 'application/json',
       }),
       body: (params) => {
-        if (!params.file || typeof params.file !== 'object') {
-          throw new Error('Identity document is required')
-        }
-
         const requestBody: Record<string, unknown> = {
           accessKeyId: params.accessKeyId?.trim(),
           secretAccessKey: params.secretAccessKey?.trim(),
           region: params.region?.trim(),
-          file: params.file,
+        }
+
+        if (params.file && typeof params.file === 'object') {
+          requestBody.file = params.file
+        } else if (params.filePath && params.filePath.trim() !== '') {
+          requestBody.filePath = params.filePath.trim()
+        } else {
+          throw new Error('Identity document is required')
         }
 
         if (params.fileBack && typeof params.fileBack === 'object') {
           requestBody.fileBack = params.fileBack
+        } else if (params.filePathBack && params.filePathBack.trim() !== '') {
+          requestBody.filePathBack = params.filePathBack.trim()
         }
 
         return requestBody

apps/sim/tools/textract/textract.test.ts

@@ -114,14 +114,25 @@ describe('textract_parser_v2', () => {
 describe('textract_analyze_expense', () => {
   const body = textractAnalyzeExpenseTool.request.body!
 
-  it('throws when no file is provided for sync mode', () => {
+  it('throws when neither file nor filePath is provided for sync mode', () => {
     expect(() =>
       body({
         accessKeyId: 'key',
         secretAccessKey: 'secret',
         region: 'us-east-1',
       } as never)
-    ).toThrow('Document file is required for single-page processing')
+    ).toThrow('Document is required for single-page processing')
+  })
+
+  it('falls back to filePath when no file object is provided', () => {
+    expect(
+      body({
+        accessKeyId: 'key',
+        secretAccessKey: 'secret',
+        region: 'us-east-1',
+        filePath: ' https://example.com/receipt.pdf ',
+      } as never)
+    ).toMatchObject({ processingMode: 'sync', filePath: 'https://example.com/receipt.pdf' })
   })
 
   it('builds an async body requiring s3Uri', () => {
@@ -191,6 +202,24 @@ describe('textract_analyze_id', () => {
     })
   })
 
+  it('falls back to filePath/filePathBack when no file objects are provided', () => {
+    expect(
+      body({
+        accessKeyId: 'key',
+        secretAccessKey: 'secret',
+        region: 'us-east-1',
+        filePath: ' https://example.com/id-front.png ',
+        filePathBack: ' https://example.com/id-back.png ',
+      } as never)
+    ).toEqual({
+      accessKeyId: 'key',
+      secretAccessKey: 'secret',
+      region: 'us-east-1',
+      filePath: 'https://example.com/id-front.png',
+      filePathBack: 'https://example.com/id-back.png',
+    })
+  })
+
   it('normalizes identity documents from the response', async () => {
     const result = await textractAnalyzeIdTool.transformResponse!(
       respond({

apps/sim/tools/textract/types.ts

@@ -111,6 +111,7 @@ export interface TextractAnalyzeExpenseV2Input {
   region: string
   processingMode?: TextractProcessingMode
   file?: UserFile
+  filePath?: string
   s3Uri?: string
 }
 
@@ -171,7 +172,9 @@ export interface TextractAnalyzeIdV2Input {
   secretAccessKey: string
   region: string
   file?: UserFile
+  filePath?: string
   fileBack?: UserFile
+  filePathBack?: string
 }
 
 interface TextractIdFieldValue {