File Analysis API

The File Analysis API brings AI understanding to your documents, images, and text files. Instead of just extracting raw content, our system comprehends meaning, answers questions, and generates insights across multiple file formats using state-of-the-art models from OpenAI and Anthropic.

Supported File Types

Documents

• PDF (.pdf) - Advanced text extraction with layout preservation
• Word (.docx) - Structure-aware content processing
• Excel (.xlsx) - Intelligent spreadsheet analysis and data interpretation

Images

• PNG, JPEG - Visual content analysis using advanced vision models
• Text extraction from screenshots and scanned documents
• Chart and diagram interpretation

Text Files

• Code files (.js, .py, .java, etc.) - Syntax-aware analysis and review
• Markup (.html, .md, .xml) - Structure-preserving processing
• Configuration (.json, .yml, .csv) - Format-specific parsing

Basic Usage

Document Analysis

  const formData = new FormData();
  formData.append('payload', JSON.stringify({
    messages: [
      {
        role: 'user',
        content: 'What are the key terms and any concerning clauses in 
  this contract?'
      }
    ],
    model: 'claude-3-5-sonnet',
    max_tokens: 2000
  }));
  formData.append('file', contractFile);

  const response = await
  fetch('https://connectapi.answerr.ai/functions/v1/chat', {
    method: 'POST',
    headers: {
      'Authorization': `Bearer ${YOUR_API_KEY}`
    },
    body: formData
  });

Multi-File Analysis

  const formData = new FormData();
  formData.append('payload', JSON.stringify({
    messages: [
      {
        role: 'user',
        content: 'Compare the financial performance between Q1 and Q2 
  reports'
      }
    ],
    model: 'gpt-4o',
    max_tokens: 3000
  }));
  formData.append('file', q1ReportFile); 
  formData.append('file', q2ReportFile);

  const response = await
  fetch('https://connectapi.answerr.ai/functions/v1/chat', {
    method: 'POST',
    headers: {
      'Authorization': `Bearer ${YOUR_API_KEY}`
    },
    body: formData
  });

Processing Intelligence

Automatic Type Detection

The API intelligently routes each file to specialized processors:

• Images → Vision-enabled AI analysis
• Documents → Format-specific extraction with context preservation
• Text files → Syntax-aware parsing and semantic understanding

Content Extraction

• PDFs: Layout-aware text extraction maintaining document structure
• Word docs: Formatting context preservation with clean text rendering
• Spreadsheets: Smart table recognition with header identification
• Images: OCR combined with visual understanding for comprehensive analysis

AI Integration

Processed content integrates seamlessly with our model ecosystem. Each model brings unique strengths—OpenAI excels at technical documents and code, while Claude provides superior analysis of complex legal and business documents.

Advanced Features

Compare Mode with Files

Compare how different models interpret the same documents:

{
  "model": "gpt-4o",
  "compareModel": "claude-3-5-sonnet",
  "files": [{"url": "technical_spec.pdf", "mimeType": "application/pdf"}],
  "input": [{"role": "user", "content": "Extract the technical requirements"}]
}

Conversation Context

Build on previous file analysis with follow-up questions:

// Initial analysis
const firstResponse = await analyzeFile(document, "Summarize this research paper");

// Follow-up questions maintain context
const followUp = await analyzeFile(document, "What methodology did they use?");
const deeper = await analyzeFile(document, "How does this compare to industry standards?");

Business Applications

Legal Document Review

Law firms use File Analysis to extract key clauses, identify risks, compare contracts against standards, and answer specific questions about legal documents without manual page-by-page review.

Financial Analysis

Analyze earnings reports, financial statements, and market research. Extract projections, compare quarterly results, identify trends, and generate executive summaries from dense financial documentation.

Technical Documentation

Engineering teams query complex specifications, extract implementation requirements, analyze code repositories, and compare different versions of technical documents to identify changes and implications.

Research and Academia

Researchers analyze academic papers, extract methodologies and findings, compare approaches across publications, and identify trends without reading every word of dense technical content.

Content Management

Marketing and communications teams analyze brand guidelines, extract key messaging, compare content strategies, and ensure consistency across large document libraries.

Processing Limits and Optimization

File Constraints

• Maximum 5 files per request for optimal processing
• Document processing: Up to 90,000 tokens (≈67,500 words)
• Text file processing: Up to 40,000 tokens (≈30,000 words)
• File size: Content-based limits ensure reliable processing

Performance Features

• Batch processing: Analyze multiple related files in a single request
• Cross-file analysis: Identify relationships and differences between documents
• Token optimization: Intelligent content prioritization within model limits
• Format preservation: Maintain critical structure and formatting context

Integration Example

class FileAnalyzer {
  constructor(apiKey) {
    this.apiKey = apiKey;
    this.baseURL = 'https://connectapi.answerr.ai/functions/v1';
  }

  async analyzeFiles(files, query, options = {}) {
    const response = await fetch(`${this.baseURL}/chat`, {
      method: 'POST',
      headers: {
        'Authorization': `Bearer ${this.apiKey}`,
        'Content-Type': 'application/json'
      },
      body: JSON.stringify({
        files,
        input: [{ role: 'user', content: query }],
        model: options.model || 'auto',
        compareModel: options.compareModel,
        includeSources: options.includeSources || false,
        userId: options.userId,
        chatId: options.chatId || this.generateChatId(),
        msgId: options.msgId || this.generateMsgId()
      })
    });

    if (!response.ok) {
      throw new Error(`Analysis failed: ${response.statusText}`);
    }

    return response.json();
  }

  async analyzeDocument(fileUrl, mimeType, fileName, query) {
    return this.analyzeFiles([{
      url: fileUrl,
      mimeType,
      name: fileName
    }], query);
  }
}

// Usage examples
const analyzer = new FileAnalyzer(process.env.ANSWERR_API_KEY);

// Single document analysis
const contractAnalysis = await analyzer.analyzeDocument(
  'https://storage.com/contract.pdf',
  'application/pdf',
  'service_contract.pdf',
  'What are the termination clauses and notice requirements?'
);

// Multi-file comparison
const quarterlyComparison = await analyzer.analyzeFiles([
  { url: 'q3_report.pdf', mimeType: 'application/pdf', name: 'Q3_Report.pdf' },
  { url: 'q4_report.pdf', mimeType: 'application/pdf', name: 'Q4_Report.pdf' }
], 'Compare revenue growth and identify key performance changes');

Next Steps

File Analysis works seamlessly with all AnswerrAI features:

• Compare Mode - Compare how different models interpret your documents
• Auto-Mode - Intelligent model selection for optimal file processing
• Sources & Suggestions - Enhanced analysis with supporting context

Start analyzing files today by uploading your first document and asking questions about its contents. Transform static files into interactive knowledge sources with the File Analysis API.