cleanup

Author: waleedlatif1

apps/sim/app/api/tools/textract/parse/route.ts

@@ -496,7 +496,10 @@ export async function POST(request: NextRequest) {
       }
     }
 
-    const { bytes } = await fetchDocumentBytes(fileUrl)
+    const { bytes, contentType } = await fetchDocumentBytes(fileUrl)
+
+    // Track if this is a PDF for better error messaging
+    const isPdf = contentType.includes('pdf') || fileUrl.toLowerCase().endsWith('.pdf')
 
     const uri = '/'
 
@@ -559,15 +562,27 @@ export async function POST(request: NextRequest) {
       logger.error(`[${requestId}] Textract API error:`, errorText)
 
       let errorMessage = `Textract API error: ${textractResponse.statusText}`
+      let isUnsupportedFormat = false
       try {
         const errorJson = JSON.parse(errorText)
         if (errorJson.Message) {
           errorMessage = errorJson.Message
         } else if (errorJson.__type) {
           errorMessage = `${errorJson.__type}: ${errorJson.message || errorText}`
         }
+        // Check for unsupported document format error
+        isUnsupportedFormat =
+          errorJson.__type === 'UnsupportedDocumentException' ||
+          errorJson.Message?.toLowerCase().includes('unsupported document') ||
+          errorText.toLowerCase().includes('unsupported document')
       } catch {
-        // Use default error message
+        isUnsupportedFormat = errorText.toLowerCase().includes('unsupported document')
+      }
+
+      // Provide helpful message for unsupported format (likely multi-page PDF)
+      if (isUnsupportedFormat && isPdf) {
+        errorMessage =
+          'This document format is not supported in Single Page mode. If this is a multi-page PDF, please use "Multi-Page (PDF, TIFF via S3)" mode instead, which requires uploading your document to S3 first. Single Page mode only supports JPEG, PNG, and single-page PDF files.'
       }
 
       return NextResponse.json(

apps/sim/app/workspace/[workspaceId]/w/[workflowId]/components/chat/components/output-select/output-select.tsx

@@ -129,10 +129,6 @@ export function OutputSelect({
           ? baselineWorkflow.blocks?.[block.id]?.subBlocks?.responseFormat?.value
           : subBlockValues?.[block.id]?.responseFormat
       const responseFormat = parseResponseFormatSafely(responseFormatValue, block.id)
-      const operationValue =
-        shouldUseBaseline && baselineWorkflow
-          ? baselineWorkflow.blocks?.[block.id]?.subBlocks?.operation?.value
-          : subBlockValues?.[block.id]?.operation
 
       let outputsToProcess: Record<string, unknown> = {}
 

apps/sim/app/workspace/[workspaceId]/w/[workflowId]/components/panel/components/editor/components/sub-block/components/tag-dropdown/tag-dropdown.tsx

@@ -242,7 +242,6 @@ const getOutputTypeForPath = (
     const subBlocks = mergedSubBlocksOverride ?? (blockState?.subBlocks || {})
     return getBlockOutputType(block.type, outputPath, subBlocks)
   } else if (blockConfig) {
-    // Pass full subBlocks to support tool selectors that use any field (operation, provider, etc.)
     const blockState = useWorkflowStore.getState().blocks[blockId]
     const subBlocks = mergedSubBlocksOverride ?? (blockState?.subBlocks || {})
     return getToolOutputType(blockConfig, subBlocks, outputPath)
@@ -1211,7 +1210,6 @@ export const TagDropdown: React.FC<TagDropdownProps> = ({
               : allTags
           }
         } else {
-          // Pass full subBlocks to support tool selectors that use any field (operation, provider, etc.)
           const toolOutputPaths = getToolOutputPaths(blockConfig, mergedSubBlocks)
 
           if (toolOutputPaths.length > 0) {

apps/sim/blocks/blocks/file.ts

@@ -150,7 +150,7 @@ export const FileV2Block: BlockConfig<FileParserOutput> = {
       title: 'Files',
       type: 'short-input' as SubBlockType,
       canonicalParamId: 'fileInput',
-      placeholder: 'File URL or reference from previous block',
+      placeholder: 'File URL',
       mode: 'advanced',
     },
   ],

apps/sim/blocks/blocks/mistral_parse.ts

@@ -165,7 +165,7 @@ export const MistralParseV2Block: BlockConfig<MistralParserOutput> = {
       title: 'PDF Document',
       type: 'short-input' as SubBlockType,
       canonicalParamId: 'document',
-      placeholder: 'Document URL or reference from previous block',
+      placeholder: 'Document URL',
       mode: 'advanced',
     },
     {

apps/sim/blocks/blocks/pulse.ts

@@ -29,7 +29,7 @@ export const PulseBlock: BlockConfig<PulseParserOutput> = {
       title: 'Document',
       type: 'short-input' as SubBlockType,
       canonicalParamId: 'document',
-      placeholder: 'Document URL or reference from previous block',
+      placeholder: 'Document URL',
       mode: 'advanced',
     },
     {

apps/sim/blocks/blocks/reducto.ts

@@ -28,7 +28,7 @@ export const ReductoBlock: BlockConfig<ReductoParserOutput> = {
       title: 'PDF Document',
       type: 'short-input' as SubBlockType,
       canonicalParamId: 'document',
-      placeholder: 'Document URL or reference from previous block',
+      placeholder: 'Document URL',
       mode: 'advanced',
     },
     {

apps/sim/blocks/blocks/textract.ts

@@ -15,20 +15,22 @@ export const TextractBlock: BlockConfig<TextractParserOutput> = {
   subBlocks: [
     {
       id: 'processingMode',
-      title: 'Document Type',
+      title: 'Processing Mode',
       type: 'dropdown' as SubBlockType,
       options: [
-        { id: 'sync', label: 'Single Page' },
-        { id: 'async', label: 'Multi-Page' },
+        { id: 'sync', label: 'Single Page (JPEG, PNG, 1-page PDF)' },
+        { id: 'async', label: 'Multi-Page (PDF, TIFF via S3)' },
       ],
+      tooltip:
+        'Single Page uses synchronous API for JPEG, PNG, or single-page PDF. Multi-Page uses async API for multi-page PDF/TIFF stored in S3.',
     },
     {
       id: 'fileUpload',
       title: 'Document',
       type: 'file-upload' as SubBlockType,
       canonicalParamId: 'document',
-      acceptedTypes: 'application/pdf,image/jpeg,image/png,image/tiff',
-      placeholder: 'Upload a document',
+      acceptedTypes: 'image/jpeg,image/png,application/pdf',
+      placeholder: 'Upload JPEG, PNG, or single-page PDF (max 10MB)',
       condition: {
         field: 'processingMode',
         value: 'async',
@@ -42,7 +44,7 @@ export const TextractBlock: BlockConfig<TextractParserOutput> = {
       title: 'Document',
       type: 'short-input' as SubBlockType,
       canonicalParamId: 'document',
-      placeholder: 'Document URL or reference from previous block',
+      placeholder: 'URL to JPEG, PNG, or single-page PDF',
       condition: {
         field: 'processingMode',
         value: 'async',

apps/sim/tools/mistral/types.ts

@@ -1,176 +1,80 @@
 import type { ToolResponse } from '@/tools/types'
 
-/**
- * Input parameters for the Mistral OCR parser tool
- */
 export interface MistralParserInput {
-  /** URL to a PDF document to be processed */
   filePath: string
-
-  /** File upload data (from file-upload component) */
   fileUpload?: any
-
-  /** Internal file path flag (for presigned URL conversion) */
   _internalFilePath?: string
-
-  /** Mistral API key for authentication */
   apiKey: string
-
-  /** Output format for the extracted content (default: 'markdown') */
   resultType?: 'markdown' | 'text' | 'json'
-
-  /** Whether to include base64-encoded images in the response */
   includeImageBase64?: boolean
-
-  /** Specific pages to process (zero-indexed) */
   pages?: number[]
-
-  /** Maximum number of images to extract from the PDF */
   imageLimit?: number
-
-  /** Minimum height and width (in pixels) for images to extract */
   imageMinSize?: number
 }
 
-/**
- * Usage information returned by the Mistral OCR API
- */
 export interface MistralOcrUsageInfo {
-  /** Number of pages processed in the document */
   pagesProcessed: number
-
-  /** Size of the document in bytes */
   docSizeBytes: number | null
 }
 
-/**
- * Metadata about the processed document
- */
 export interface MistralParserMetadata {
-  /** Unique identifier for this OCR job */
   jobId: string
-
-  /** File type of the document (typically 'pdf') */
   fileType: string
-
-  /** Filename extracted from the document URL */
   fileName: string
-
-  /** Source type (always 'url' for now) */
   source: 'url'
-
-  /** Original URL to the document (only included for user-provided URLs) */
   sourceUrl?: string
-
-  /** Total number of pages in the document */
   pageCount: number
-
-  /** Usage statistics from the OCR processing */
   usageInfo?: MistralOcrUsageInfo
-
-  /** The Mistral OCR model used for processing */
   model: string
-
-  /** The output format that was requested */
   resultType?: 'markdown' | 'text' | 'json'
-
-  /** ISO timestamp when the document was processed */
   processedAt: string
 }
 
-/**
- * Output data structure from the Mistral OCR parser
- */
 export interface MistralParserOutputData {
-  /** Extracted content in the requested format */
   content: string
-
-  /** Metadata about the parsed document and processing */
   metadata: MistralParserMetadata
 }
 
-/**
- * Complete response from the Mistral OCR parser tool
- */
 export interface MistralParserOutput extends ToolResponse {
-  /** The output data containing content and metadata */
   output: MistralParserOutputData
 }
 
-/**
- * Image bounding box and data from Mistral OCR API
- */
 export interface MistralOcrImage {
-  /** Image identifier */
   id: string
-  /** Top-left X coordinate */
   top_left_x: number
-  /** Top-left Y coordinate */
   top_left_y: number
-  /** Bottom-right X coordinate */
   bottom_right_x: number
-  /** Bottom-right Y coordinate */
   bottom_right_y: number
-  /** Base64-encoded image data (if includeImageBase64 was true) */
   image_base64?: string
 }
 
-/**
- * Page dimensions from Mistral OCR API
- */
 export interface MistralOcrDimensions {
-  /** DPI of the page */
   dpi: number
-  /** Page height in pixels */
   height: number
-  /** Page width in pixels */
   width: number
 }
 
-/**
- * Page data from Mistral OCR API
- */
 export interface MistralOcrPage {
-  /** Page index (zero-based) */
   index: number
-  /** Markdown content extracted from this page */
   markdown: string
-  /** Images extracted from this page */
   images: MistralOcrImage[]
-  /** Page dimensions */
   dimensions: MistralOcrDimensions
-  /** Tables extracted from this page */
   tables: unknown[]
-  /** Hyperlinks found on this page */
   hyperlinks: unknown[]
-  /** Header content if detected */
   header: string | null
-  /** Footer content if detected */
   footer: string | null
 }
 
-/**
- * Raw usage info from Mistral OCR API
- */
 export interface MistralOcrUsageInfoRaw {
-  /** Number of pages processed */
   pages_processed: number
-  /** Document size in bytes */
   doc_size_bytes: number | null
 }
 
-/**
- * V2 Output - Returns raw Mistral API response structure
- */
 export interface MistralParserV2Output extends ToolResponse {
   output: {
-    /** Array of page objects with full OCR data */
     pages: MistralOcrPage[]
-    /** Model used for OCR processing */
     model: string
-    /** Usage statistics from the API */
     usage_info: MistralOcrUsageInfoRaw
-    /** Structured annotation data as JSON string (when applicable) */
     document_annotation: string | null
   }
 }