File Uploads

Handle file uploads with paths and file-like objects.

📁 Upload Files by Path

Single File

from glaip_sdk import Client

client = Client()
agent = client.create_agent(
    name="file-processor",
    instruction="Process uploaded files and provide insights."
)

# Upload file by path
response = agent.run(
    "Analyze this document",
    files=["document.pdf"]
)

Multiple Files

# Upload multiple files
response = agent.run(
    "Compare these documents",
    files=["doc1.pdf", "doc2.pdf", "doc3.pdf"]
)

CLI File Format: Use repeatable --file flags (--file "a" --file "b" --file "c"). Each file is specified with a separate --file flag.

📂 Upload File-Like Objects

File Handles

# Upload file-like objects
with open("resume.pdf", "rb") as f:
    response = agent.run(
        "Summarize this resume",
        files=[f]
    )

Bytes and IO Objects

import io

# Upload from bytes
file_content = b"Hello, this is file content"
file_obj = io.BytesIO(file_content)
file_obj.name = "example.txt"

response = agent.run(
    "Process this text file",
    files=[file_obj]
)

# Upload from BytesIO
import io
text_content = "This is text content"
text_obj = io.BytesIO(text_content.encode("utf-8"))
text_obj.name = "text.txt"

response = agent.run(
    "Analyze this text",
    files=[text_obj]
)

🔧 File Upload Configuration

Supported File Types

File Type
Extensions
Description

Documents

.pdf, .docx, .txt, .md

Text and document files

Images

.jpg, .png, .gif, .svg

Image files for analysis

Data

.csv, .json, .xml

Structured data files

Code

.py, .js, .java, .cpp

Source code files

Archives

.zip, .tar.gz

Compressed files

File Size Limits

  • Maximum file size: 100 MB per file (default)

  • Total upload limit: 500 MB per request (default)

  • File count limit: 10 files per request (default)

Note: These are default limits that may be configured differently on your server. Check with your system administrator for specific limits.

Large files may take longer to process. Consider compressing files when possible.

Shell Considerations

Glob Expansion: On Linux/macOS, shell glob expansion works automatically (e.g., *.pdf). On Windows PowerShell, use explicit file lists instead of glob patterns.

📊 File Processing Examples

Document Analysis

# Analyze PDF document
response = agent.run(
    "Extract key information from this document",
    files=["report.pdf"]
)

# Process multiple documents
response = agent.run(
    "Compare the main points in these reports",
    files=["report1.pdf", "report2.pdf", "report3.pdf"]
)

Image Analysis

# Analyze image content
response = agent.run(
    "Describe what you see in this image",
    files=["photo.jpg"]
)

# Compare multiple images
response = agent.run(
    "Find differences between these images",
    files=["image1.png", "image2.png"]
)

Data Analysis

# Analyze CSV data
response = agent.run(
    "Summarize the trends in this dataset",
    files=["sales_data.csv"]
)

# Process JSON configuration
response = agent.run(
    "Validate this configuration file",
    files=["config.json"]
)

Code Review

# Review Python code
response = agent.run(
    "Review this code for best practices",
    files=["main.py"]
)

# Analyze multiple source files
response = agent.run(
    "Find potential issues in this codebase",
    files=["main.py", "utils.py", "tests.py"]
)

🔄 Advanced File Handling

Dynamic File Selection

import os
import glob

# Upload all PDF files in a directory
pdf_files = glob.glob("documents/*.pdf")
response = agent.run(
    "Process all PDF documents",
    files=pdf_files
)

# Upload files based on pattern
txt_files = glob.glob("data/*.txt")
response = agent.run(
    "Analyze all text files",
    files=txt_files
)

File Validation

import os

def validate_files(file_paths):
    """Validate files before upload."""
    valid_files = []
    for path in file_paths:
        if os.path.exists(path):
            file_size = os.path.getsize(path)
            if file_size <= 100 * 1024 * 1024:  # 100 MB
                valid_files.append(path)
            else:
                print(f"File too large: {path}")
        else:
            print(f"File not found: {path}")
    return valid_files

# Validate and upload
file_paths = ["doc1.pdf", "doc2.pdf", "doc3.pdf"]
valid_files = validate_files(file_paths)

if valid_files:
    response = agent.run(
        "Process valid documents",
        files=valid_files
    )

File Limits

# File size validation
import os

def check_file_size(file_path):
    file_size = os.path.getsize(file_path)
    max_size = 100 * 1024 * 1024  # 100 MB
    return file_size <= max_size

# Check before upload
if check_file_size("large_file.pdf"):
    response = agent.run("Process file", files=["large_file.pdf"])
else:
    print("File too large")

Note: File size limits (100 MB) are server defaults and may vary by deployment. Check with your AIP administrator for specific limits.

🎯 Best Practices

File Preparation

  • Compress large files: Use ZIP or TAR for multiple files

  • Validate file types: Ensure files are in supported formats

  • Check file sizes: Stay within upload limits

  • Organize files: Use descriptive filenames and organize in directories

Upload Strategy

  • Batch processing: Upload related files together

  • Progressive uploads: Start with small files to test

  • Error handling: Handle upload failures gracefully

  • Cleanup: Remove temporary files after processing

Performance Optimization

  • File size: Keep files under 50 MB when possible

  • File count: Limit to 5-10 files per request

  • Format selection: Use efficient formats (e.g., CSV over Excel)

  • Compression: Compress text-based files

🚀 Next Steps

  • Tools: Learn about Tools for processing uploaded files

  • MCPs: Connect external services via MCPs

  • Streaming: Use Streaming & Renderers for real-time file processing

  • Scripting: Automate file workflows with Output & Scripting

💡 Pro Tips

  • Start Small: Test with small files before uploading large ones

  • Use Paths: File paths are simpler than file objects for most use cases

  • Validate First: Check file existence and size before upload

  • Organize Files: Keep related files in organized directories

  • Monitor Progress: Large files may take time to process

  • Handle Errors: Implement proper error handling for file operations

Ready to handle file uploads? Start with simple documents and gradually work with more complex file types!

PreviousMCPsNextStreaming & Renderers