📄 Document Processing API v1.2.3

GET /health

Checks the operational status and version of the API service.

200 OK Service is healthy.

{
    "status": "healthy",
    "version": "1.2.3",
    "timestamp": "YYYY-MM-DDTHH:MM:SSZ",
    "service": "document-processing-api"
}

POST /document-info

Identifies the type of the document, lists key information categories, and suggests custom prompt templates for detailed extraction with the /summarize endpoint. File content must be base64 encoded.

Request Body (application/json)

Parameter	Type	Required	Description
fileName	string	Yes	Name of the file (e.g., "invoice.png"). Used for MIME type detection.
fileContent	string	Yes	Base64 encoded string of the file content.

Example Request

{
    "fileName": "statement_may_2023.pdf",
    "fileContent": "BASE64_ENCODED_FILE_CONTENT_STRING"
}

Successful Response 200 OK

{
    "document_type": "Medical Bill",
    "key_information_categories": [
        "Patient Information",
        "Provider Details",
        "Service Dates & Descriptions",
        "Charges and Payments"
    ],
    "suggested_custom_prompts": [
        {
            "prompt_name": "Basic Bill Summary",
            "description": "Extract key summary details from the medical bill.",
            "prompt_template_for_user": "Please extract the following basic details from this medical bill. Respond with a single, valid JSON object only. The JSON structure should be: {\\\"patient_name\\\": \\\"\\\", \\\"provider_name\\\": \\\"\\\", \\\"total_charges\\\": \\\"\\\", \\\"amount_due\\\": \\\"\\\"}"
        },
        {
            "prompt_name": "Detailed Charges Extraction",
            "description": "Extract a comprehensive list of all itemized charges from the medical bill.",
            "prompt_template_for_user": "Extract all itemized charges from this medical bill. Your response must be a single, valid JSON object. The JSON structure should be: {\\\"bill_id\\\": \\\"\\\", \\\"patient_id\\\": \\\"\\\", \\\"itemized_charges\\\": [{\\\"service_date\\\": \\\"\\\", \\\"service_description\\\": \\\"\\\", \\\"cpt_code\\\": \\\"\\\", \\\"charge_amount\\\": \\\"\\\"}], \\\"total_billed_amount\\\": \\\"\\\"}"
        },
        {
            "prompt_name": "Insurance Payment Information",
            "description": "Find details about any insurance payments or adjustments mentioned on the bill.",
            "prompt_template_for_user": "From this medical bill, if insurance payment details are present, extract them. Otherwise, indicate no insurance payment found. Respond in JSON. The JSON structure should be: {\\\"insurance_payment_found\\\": , \\\"insurance_provider\\\": \\\"\\\", \\\"payment_amount\\\": \\\"\\\", \\\"adjustment_amount\\\": \\\"\\\"}"
        }
    ]
}

Note: The suggested_custom_prompts array provides different templates. Choose and adapt one of the prompt_template_for_user strings for the customJsonPrompt in a POST /summarize request.

POST /summarize

Analyzes and summarizes a document. If customJsonPrompt is provided, it attempts to extract data according to that prompt. Otherwise, it provides a general summary. The file content must be base64 encoded.

Request Body (application/json)

Parameter	Type	Required	Description
fileName	string	Yes	Name of the file (e.g., "report.pdf"). Used for MIME type detection and S3 naming.
fileContent	string	Yes	Base64 encoded string of the file content.
customJsonPrompt	string	No	Your specific instructions for the AI, expecting a JSON output. If provided, the API returns the raw JSON from the AI. You can use the /document-info endpoint to help generate this.
summaryLength	string	No	Desired summary length: "short", "medium", or "long". Defaults to "medium". Only used if customJsonPrompt is empty.

Example Request (Default Summary)

{
    "fileName": "annual_report.docx",
    "fileContent": "BASE64_ENCODED_FILE_CONTENT_STRING",
    "summaryLength": "medium"
}

Successful Response (Default Summary) 200 OK

{
    "title": "Document Title",
    "key_points": ["Key point 1", "Key point 2"],
    "overall_summary": "A narrative summary...",
    "estimated_reading_time_minutes": 5,
    "sentiment": "Positive"
}

Example Request (Custom JSON Extraction)

{
    "fileName": "invoice_123.pdf",
    "fileContent": "BASE64_ENCODED_FILE_CONTENT_STRING",
    "customJsonPrompt": "Please extract the following details from this invoice and return your response as a single, valid JSON object only. {\"invoice_id\": \"\", \"vendor_name\": \"\", \"total_due\": \"\"}"
}

Successful Response (Custom JSON Extraction) 200 OK

Returns the raw JSON object generated by the AI based on your customJsonPrompt. For example:

{
    "invoice_id": "INV-2023-001",
    "vendor_name": "Supplier Corp",
    "total_due": "1250.75"
}