Python SDK - Daytona

The Daytona Python SDK provides a simple interface for interacting with Daytona Sandboxes, enabling sandbox management, code execution, file operations, and more.

Installation

pip install daytona

Quick Start

from daytona import Daytona

# Initialize using environment variables
daytona = Daytona()

# Create a sandbox
sandbox = daytona.create()

# Run code in the sandbox
response = sandbox.process.code_run('print("Hello World!")')
print(response.result)

# Clean up when done
daytona.delete(sandbox)

Configuration

Using Environment Variables

Set the following environment variables:

export DAYTONA_API_KEY="your-api-key"
export DAYTONA_API_URL="https://app.daytona.io/api"  # Optional
export DAYTONA_TARGET="us"  # Optional

Then initialize the client:

from daytona import Daytona

daytona = Daytona()

Using Configuration Object

from daytona import Daytona, DaytonaConfig

config = DaytonaConfig(
    api_key="your-api-key",
    api_url="https://app.daytona.io/api",
    target="us"
)
daytona = Daytona(config)

Sandbox Management

Creating Sandboxes

Basic Sandbox Creation

from daytona import Daytona

daytona = Daytona()
sandbox = daytona.create()

print(f"Created sandbox: {sandbox.id}")
print(f"State: {sandbox.state}")

Customized Sandbox Creation

from daytona import Daytona, CreateSandboxFromSnapshotParams

daytona = Daytona()

params = CreateSandboxFromSnapshotParams(
    language="python",
    env_vars={"PYTHON_ENV": "development"},
    auto_stop_interval=60,  # Auto-stop after 1 hour of inactivity
    auto_archive_interval=60,  # Auto-archive after 1 hour of being stopped
    auto_delete_interval=120  # Auto-delete after 2 hours of being stopped
)

sandbox = daytona.create(params)

Listing Sandboxes

# List all sandboxes
result = daytona.list()
print(f"Total sandboxes: {result.total}")

for sandbox in result.items:
    print(f"ID: {sandbox.id}, State: {sandbox.state}")

# List with pagination
result = daytona.list(page=1, limit=10)

Getting a Sandbox

# Get by ID
sandbox = daytona.get("sandbox-id")

# Get by name
sandbox = daytona.get("sandbox-name")

Sandbox Lifecycle

# Stop a sandbox
daytona.stop(sandbox)
print(f"Sandbox state: {sandbox.state}")  # 'stopped'

# Start a sandbox
daytona.start(sandbox)
print(f"Sandbox state: {sandbox.state}")  # 'started'

# Delete a sandbox
daytona.delete(sandbox)

Setting Labels

sandbox.set_labels({
    "environment": "development",
    "project": "my-app",
    "public": "true"
})

print(f"Labels: {sandbox.labels}")

Code Execution

Running Python Code

# Simple code execution
response = sandbox.process.code_run('print("Hello, World!")')
print(response.result)

# Multi-line code
code = '''
x = 10
y = 20
print(f"Sum: {x + y}")
'''

response = sandbox.process.code_run(code)
print(response.result)  # Output: Sum: 30

Executing Shell Commands

# Execute a command
response = sandbox.process.exec('echo "Hello, World!"')
if response.exit_code == 0:
    print(response.result)
else:
    print(f"Error: {response.result}")

# Execute with working directory and timeout
response = sandbox.process.exec(
    'ls -la',
    cwd='/home/daytona',
    timeout=10
)
print(response.result)

Code Interpreter (Stateful Execution)

# Create a stateful code interpreter session
interpreter = sandbox.code_interpreter

# Execute code with real-time streaming
channels = interpreter.run_code('''
import time
for i in range(5):
    print(f"Step {i+1}")
    time.sleep(0.5)
''')

# Read real-time output
for msg in channels.stdout:
    print(f"[STDOUT] {msg.text}")

# Wait for completion
result = next(channels.done)
print(f"Final output: {result.stdout}")

File Operations

Uploading Files

# Upload a single file from bytes
content = b'Hello, World!'
sandbox.fs.upload_file(content, '/tmp/hello.txt')

# Upload a file from local path
from daytona import FileUpload

sandbox.fs.upload_files([
    FileUpload(source='local-file.txt', destination='/tmp/remote-file.txt')
])

# Upload multiple files
sandbox.fs.upload_files([
    FileUpload(source='file1.txt', destination='/tmp/file1.txt'),
    FileUpload(source=b'content', destination='/tmp/file2.txt')
])

Downloading Files

# Download a single file to memory
content = sandbox.fs.download_file('/tmp/hello.txt')
print(content.decode('utf-8'))

# Download multiple files
from daytona import FileDownloadRequest

results = sandbox.fs.download_files([
    FileDownloadRequest(source='/tmp/file1.txt', destination='local-file1.txt'),
    FileDownloadRequest(source='/tmp/file2.txt')  # To memory
])

for result in results:
    if result.error:
        print(f"Error: {result.error}")
    elif isinstance(result.result, str):
        print(f"Downloaded to: {result.result}")
    else:
        print(f"Downloaded to memory: {len(result.result)} bytes")

File System Operations

# List files in a directory
files = sandbox.fs.list_files('/tmp')
for file in files:
    print(f"{file.name}: {file.size} bytes")

# Create a folder
sandbox.fs.create_folder('/tmp/new-folder', mode='755')

# Search for files
matches = sandbox.fs.search_files('/tmp', '*.txt')
print(f"Found {len(matches.files)} files")
for file in matches.files:
    print(file)

# Replace content in files
sandbox.fs.replace_in_files(
    ['/tmp/config.json'],
    '"debug": true',
    '"debug": false'
)

Git Operations

Cloning Repositories

# Clone a public repository
sandbox.git.clone(
    'https://github.com/daytonaio/daytona.git',
    '/tmp/daytona'
)

# Clone with authentication
sandbox.git.clone(
    'https://github.com/private/repo.git',
    '/tmp/repo',
    username='user',
    password='token'
)

# Clone a specific branch
sandbox.git.clone(
    'https://github.com/example/repo.git',
    '/tmp/repo',
    branch='develop'
)

Git Status and Branches

# Get repository status
status = sandbox.git.status('/tmp/repo')
print(f"Current branch: {status.current_branch}")
print(f"Files changed: {len(status.file_status)}")

# List branches
branches = sandbox.git.branches('/tmp/repo')
for branch in branches:
    print(f"Branch: {branch}")

Committing Changes

# Add files
sandbox.git.add('/tmp/repo', ['file1.txt', 'file2.txt'])

# Commit changes
sandbox.git.commit(
    '/tmp/repo',
    'Add new features',
    'Your Name',
    '[email protected]'
)

# Push changes
sandbox.git.push(
    '/tmp/repo',
    username='user',
    password='token'
)

Language Server Protocol

Starting an LSP Server

# Create and start a TypeScript LSP server
lsp = sandbox.create_lsp_server('typescript', '/workspace/project')
lsp.start()

# Notify LSP about opened file
lsp.did_open('/workspace/project/src/index.ts')

Code Intelligence

# Get document symbols
symbols = lsp.document_symbols('/workspace/project/src/index.ts')
for symbol in symbols:
    print(f"{symbol.name} ({symbol.kind})")

# Get completions
completions = lsp.completions(
    '/workspace/project/src/index.ts',
    {"line": 10, "character": 15}
)
for item in completions:
    print(f"{item.label}: {item.detail}")

# Go to definition
definition = lsp.definition(
    '/workspace/project/src/index.ts',
    {"line": 5, "character": 10}
)
print(f"Definition at: {definition}")

Stopping LSP Server

# Stop the LSP server
lsp.stop()

Async Support

The Python SDK provides full async support:

import asyncio
from daytona import AsyncDaytona

async def main():
    # Initialize async client
    daytona = AsyncDaytona()
    
    # Create sandbox asynchronously
    sandbox = await daytona.create()
    
    # Execute code
    response = await sandbox.process.code_run('print("Async Hello!")')
    print(response.result)
    
    # Cleanup
    await daytona.delete(sandbox)

# Run the async function
asyncio.run(main())

Error Handling

from daytona import (
    Daytona,
    DaytonaError,
    DaytonaNotFoundError,
    DaytonaRateLimitError,
    DaytonaTimeoutError
)

try:
    daytona = Daytona()
    sandbox = daytona.create()
    
    # Execute code that might fail
    response = sandbox.process.exec('invalid-command')
    if response.exit_code != 0:
        print(f"Command failed: {response.result}")
    
except DaytonaNotFoundError as e:
    print(f"Resource not found: {e}")
except DaytonaRateLimitError as e:
    print(f"Rate limit exceeded: {e}")
except DaytonaTimeoutError as e:
    print(f"Operation timed out: {e}")
except DaytonaError as e:
    print(f"Daytona error: {e}")
finally:
    # Always cleanup
    if 'sandbox' in locals():
        daytona.delete(sandbox)

Complete Example

from daytona import Daytona, CreateSandboxFromSnapshotParams

def main():
    # Initialize client
    daytona = Daytona()
    
    # Create a sandbox with custom configuration
    params = CreateSandboxFromSnapshotParams(
        language="python",
        env_vars={"ENV": "production"},
        labels={"project": "data-analysis"}
    )
    
    sandbox = daytona.create(params)
    print(f"Created sandbox: {sandbox.id}")
    
    try:
        # Upload a Python script
        script = b'''
import pandas as pd
import json

data = {'name': ['Alice', 'Bob'], 'age': [25, 30]}
df = pd.DataFrame(data)
print(json.dumps(df.to_dict()))
'''
        sandbox.fs.upload_file(script, '/tmp/analyze.py')
        
        # Install dependencies
        install_result = sandbox.process.exec('pip install pandas')
        if install_result.exit_code != 0:
            print(f"Failed to install pandas: {install_result.result}")
            return
        
        # Run the script
        result = sandbox.process.exec('python /tmp/analyze.py')
        print(f"Analysis result: {result.result}")
        
    finally:
        # Cleanup
        daytona.delete(sandbox)
        print("Sandbox deleted")

if __name__ == "__main__":
    main()

API Reference

Daytona Class

Methods

create(params=None) - Create a new sandbox
get(sandbox_id_or_name) - Get a sandbox by ID or name
list(labels=None, page=None, limit=None) - List sandboxes with pagination
delete(sandbox) - Delete a sandbox
start(sandbox) - Start a stopped sandbox
stop(sandbox) - Stop a running sandbox

Sandbox Class

Properties

fs - FileSystem operations interface
git - Git operations interface
process - Process execution interface
code_interpreter - Code interpreter interface
id - Sandbox ID
state - Current state
labels - Custom labels
env - Environment variables

Methods

set_labels(labels) - Set custom labels
get_preview_link(port) - Get port preview URL

Process Class

Methods

exec(command, cwd=None, timeout=None) - Execute a shell command
code_run(code, params=None) - Execute code directly

FileSystem Class

Methods

upload_file(content, path) - Upload a file
upload_files(files) - Upload multiple files
download_file(path) - Download a file
download_files(requests) - Download multiple files
list_files(path) - List files in a directory
create_folder(path, mode='755') - Create a folder
search_files(root, pattern) - Search for files
replace_in_files(files, old, new) - Replace content in files

Git Class

Methods

clone(url, path, branch=None, username=None, password=None) - Clone a repository
status(path) - Get repository status
branches(path) - List branches
add(path, files) - Add files to staging
commit(path, message, author, email) - Commit changes
push(path, username=None, password=None) - Push changes

Best Practices

Always Use Context Managers

from daytona import Daytona

def process_data():
    daytona = Daytona()
    sandbox = daytona.create()
    
    try:
        # Do work with sandbox
        result = sandbox.process.code_run('print("Hello")')
        return result
    finally:
        # Always cleanup
        daytona.delete(sandbox)

Check Exit Codes

response = sandbox.process.exec('some-command')
if response.exit_code != 0:
    print(f"Command failed: {response.result}")
    # Handle error
else:
    print(f"Success: {response.result}")

Use Type Hints

from daytona import Daytona, Sandbox, CreateSandboxFromSnapshotParams

def create_python_sandbox(daytona: Daytona) -> Sandbox:
    params = CreateSandboxFromSnapshotParams(language="python")
    return daytona.create(params)

Next Steps

SDK Overview

Compare all available SDKs

API Reference

Complete API documentation

Examples

Browse Python examples on GitHub

TypeScript SDK

Learn about the TypeScript SDK

Getting Started

Core Concepts

SDK Guides

Features

CLI Reference

Advanced

​Installation

​Quick Start

​Configuration

​Using Environment Variables

​Using Configuration Object

​Sandbox Management

​Creating Sandboxes

​Basic Sandbox Creation

​Customized Sandbox Creation

​Listing Sandboxes

​Getting a Sandbox

​Sandbox Lifecycle

​Setting Labels

​Code Execution

​Running Python Code

​Executing Shell Commands

​Code Interpreter (Stateful Execution)

​File Operations

​Uploading Files

​Downloading Files

​File System Operations

​Git Operations

​Cloning Repositories

​Git Status and Branches

​Committing Changes

​Language Server Protocol

​Starting an LSP Server

​Code Intelligence

​Stopping LSP Server

​Async Support

​Error Handling

​Complete Example

​API Reference

​Daytona Class

​Methods

​Sandbox Class

​Properties

​Methods

​Process Class

​Methods

​FileSystem Class

​Methods

​Git Class

​Methods

​Best Practices

​Always Use Context Managers

​Check Exit Codes

​Use Type Hints

​Next Steps

SDK Overview

API Reference

Examples

TypeScript SDK

Build docs developers (and LLMs) love

Installation

Quick Start

Configuration

Using Environment Variables

Using Configuration Object

Sandbox Management

Creating Sandboxes

Basic Sandbox Creation

Customized Sandbox Creation

Listing Sandboxes

Getting a Sandbox

Sandbox Lifecycle

Setting Labels

Code Execution

Running Python Code

Executing Shell Commands

Code Interpreter (Stateful Execution)

File Operations

Uploading Files

Downloading Files

File System Operations

Git Operations

Cloning Repositories

Git Status and Branches

Committing Changes

Language Server Protocol

Starting an LSP Server

Code Intelligence

Stopping LSP Server

Async Support

Error Handling

Complete Example

API Reference

Daytona Class

Methods

Sandbox Class

Properties

Methods

Process Class

Methods

FileSystem Class

Methods

Git Class

Methods

Best Practices

Always Use Context Managers

Check Exit Codes

Use Type Hints

Next Steps