Skip to main content

​
Gemini Document Understanding

Gemini models support understanding PDF documents and other document formats, and can output structured results in specified formats. 
POST /v1beta/models/{model}:generateContent

​
PDF Document Understanding

curl "https://crazyrouter.com/v1beta/models/gemini-2.5-flash:generateContent?key=YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      {
        "role": "user",
        "parts": [
          {"text": "Summarize the core content of this document"},
          {
            "inlineData": {
              "mimeType": "application/pdf",
              "data": "JVBERi0xLjQKMSAwIG9iago..."
            }
          }
        ]
      }
    ]
  }'

​
Formatted Output

Use responseMimeType and responseSchema to control the output format:

​
JSON Format Output

cURL
curl "https://crazyrouter.com/v1beta/models/gemini-2.5-flash:generateContent?key=YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      {
        "role": "user",
        "parts": [
          {"text": "Extract key information from this resume"},
          {
            "inlineData": {
              "mimeType": "application/pdf",
              "data": "JVBERi0xLjQK..."
            }
          }
        ]
      }
    ],
    "generationConfig": {
      "responseMimeType": "application/json",
      "responseSchema": {
        "type": "object",
        "properties": {
          "name": {"type": "string"},
          "email": {"type": "string"},
          "phone": {"type": "string"},
          "education": {
            "type": "array",
            "items": {
              "type": "object",
              "properties": {
                "school": {"type": "string"},
                "degree": {"type": "string"},
                "year": {"type": "string"}
              }
            }
          },
          "experience": {
            "type": "array",
            "items": {
              "type": "object",
              "properties": {
                "company": {"type": "string"},
                "position": {"type": "string"},
                "duration": {"type": "string"}
              }
            }
          },
          "skills": {
            "type": "array",
            "items": {"type": "string"}
          }
        }
      }
    }
  }'

​
Response

{
  "candidates": [
    {
      "content": {
        "parts": [
          {
            "text": "{\"name\":\"John Doe\",\"email\":\"john@example.com\",\"phone\":\"555-0100\",\"education\":[{\"school\":\"MIT\",\"degree\":\"M.S. Computer Science\",\"year\":\"2020\"}],\"experience\":[{\"company\":\"Tech Corp\",\"position\":\"Senior Engineer\",\"duration\":\"2020-2024\"}],\"skills\":[\"Python\",\"Go\",\"Kubernetes\"]}"
          }
        ],
        "role": "model"
      },
      "finishReason": "STOP"
    }
  ]
}

​
Table Data Extraction

Python
response = model.generate_content(
    [
        "Extract all table data from the document and return in JSON format",
        {"mime_type": "application/pdf", "data": pdf_data}
    ],
    generation_config=genai.GenerationConfig(
        response_mime_type="application/json"
    )
)

import json
tables = json.loads(response.text)

​
Document Comparison

Python
with open("v1.pdf", "rb") as f:
    v1_data = f.read()
with open("v2.pdf", "rb") as f:
    v2_data = f.read()

response = model.generate_content([
    "Compare these two versions of the document and list all changes",
    {"mime_type": "application/pdf", "data": v1_data},
    {"mime_type": "application/pdf", "data": v2_data}
])
Using responseSchema ensures the model output strictly conforms to the specified JSON structure, making it ideal for data extraction and automation scenarios.
Large PDF documents consume a significant number of tokens. It is recommended to process documents with more than 50 pages in segments.