Portfolio · AI Tools

Document Intelligence

Upload any document and define natural-language extraction prompts. A configurable AI vision model returns structured JSON with confidence scores. Integrates with Salesforce — provision a custom object and have documents processed automatically on upload. Inspired by Dan's work on MuleSoft IDP.

3 free runs per day per visitor · PDFs, Word docs, and any image format up to 10 MB
ℹ️ How it works
Document
Document preview
1 / 1
📂
Drop file here or click to browse
PDF, DOCX, ZIP, or any image format · max 10 MB
⚡ Image compressed to fit 10 MB limit
Sample documents
—
No samples for this template yet.
Extraction Prompts
🔗 —
Template:
Vision model:
Field Name Extraction Instruction
3 runs remaining today
Example output Results — your extracted fields will appear here
JSON 
[
  {
    "field": "documentType",
    "prompt": "What type of identity document is this?",
    "value": "Passport",
    "confidence": 0.99
  },
  {
    "field": "expiryDate",
    "prompt": "What is the expiry or expiration date?",
    "value": "2029-04-15",
    "confidence": 0.98
  },
  {
    "field": "country",
    "prompt": "Which country issued this document?",
    "value": "Australia",
    "confidence": 0.97
  },
  {
    "field": "idData",
    "prompt": "Extract the primary identification number.",
    "value": "PA1234567",
    "confidence": 0.95
  }
]
CSV 
field,prompt,value,confidence
documentType,"What type of identity document?",Passport,0.99
expiryDate,"What is the expiry date?",2029-04-15,0.98
country,"Which country issued this?",Australia,0.97
idData,"Extract the ID number.",PA1234567,0.95
API Example Programmatic Access
Endpoint
POST https://danscodellaro.com/api/idp/submit/your-token
cURL Example
curl -X POST \ https://danscodellaro.com/api/idp/submit/your-token \ -F "[email protected]"
Python Example
import requests with open("document.pdf", "rb") as f: r = requests.post( "https://danscodellaro.com/api/idp/submit/your-token", files={"file": f} ) print(r.json())
Response (JSON)
{ "template": "Identity Document", "token": "your-token", "file": "passport.pdf", "processed_at": "2025-01-01T12:00:00Z", "summary": "Australian passport issued to J. Smith, expires 2029.", "verbatim_text": "AUSTRALIA\nPASSPORT\nSurname: SMITH…", "fields": [ { "field": "documentType", "prompt": "What type of identity document?", "value": "Passport", "confidence": 0.99 }, { "field": "country", "prompt": "Which country issued this?", "value": "Australia", "confidence": 0.97 } ], "remaining_runs_today": 2 }
Accepts: multipart/form-data · field name: file Max file size: 10 MB Rate limit: 3 requests / IP / day
Results
JSON Copied! 

        

          
          
        

      

      

        

          CSV
          Copied!
        

        

        

          
          
        

      

    

  


  
  

    

      API
      Programmatic Access
      

        
        
      

      
    

    


      
      

        

          
Endpoint

          

            POST
            
            
          

          

        

        

          
cURL Example

          

            
            
          

          

        

        

          
Python Example

          

            
            
          

          

        

      


      
      

        

          
Submit (returns immediately)

          

            POST
            
            
          

          

        

        

          
Poll for result

          

            GET
            /api/idp/jobs/{job_id}
          

        

        

          
Python Example

          

            
            
          

          

        

      


      
      

        

          
Response (JSON)

          

            
            
          

          

        

        
      


    

    

      Accepts: multipart/form-data · field name: file
      Max file size: 10 MB
      Supported: PDF, DOCX, JPG, PNG, WEBP, TIFF, BMP, GIF, HEIC
      Rate limit: 3 requests / IP / day · Jobs expire after 1 hour
    

  















  







  
    
  





  

    Ask about Dan