Runpod | Python Library

Welcome to the official Python library for Runpod API & SDK.

Table of Contents

• Table of Contents
• 💻 | Installation
• ⚡ | Serverless Worker (SDK)
  • Quick Start
  • Local Test Worker
• 📚 | API Language Library (GraphQL Wrapper)
  • Endpoints
  • GPU Cloud (Pods)
• 📁 | Directory
• 🤝 | Community and Contributing

💻 | Installation

Install from PyPI (Stable Release)

# Install with pip
pip install runpod
# Install with uv (faster alternative)
uv add runpod

Install from GitHub (Latest Changes)

To get the latest changes that haven't been released to PyPI yet:

# Install latest development version from main branch with pip
pip install git+https://github.com/runpod/runpod-python.git
# Install with uv
uv add git+https://github.com/runpod/runpod-python.git
# Install a specific branch
pip install git+https://github.com/runpod/runpod-python.git@branch-name
# Install a specific tag/release
pip install git+https://github.com/runpod/runpod-python.git@v1.0.0
# Install in editable mode for development
git clone https://github.com/runpod/runpod-python.git
cd runpod-python
pip install -e .

Python 3.10 or higher is required to use the latest version of this package.

⚡ | Serverless Worker (SDK)

This python package can also be used to create a serverless worker that can be deployed to Runpod as a custom endpoint API.

Quick Start

Create a python script in your project that contains your model definition and the Runpod worker start code. Run this python code as your default container start command:

# my_worker.pyimportrunpoddefis_even(job):
job_input=job["input"]
the_number=job_input["number"]
ifnotisinstance(the_number, int):
return {"error": "Silly human, you need to pass an integer."}
ifthe_number%2==0:
returnTruereturnFalserunpod.serverless.start({"handler": is_even})

Make sure that this file is ran when your container starts. This can be accomplished by calling it in the docker command when you set up a template at console.runpod.io/serverless/user/templates or by setting it as the default command in your Dockerfile.

See our blog post for creating a basic Serverless API, or view the details docs for more information.

Local Test Worker

You can also test your worker locally before deploying it to Runpod. This is useful for debugging and testing.

python my_worker.py --rp_serve_api

Worker Fitness Checks

Fitness checks allow you to validate your worker environment at startup before processing jobs. If any check fails, the worker exits immediately, allowing your orchestrator to restart it.

# my_worker.pyimportrunpodimporttorch# Register fitness checks using the decorator@runpod.serverless.register_fitness_checkdefcheck_gpu_available():
"""Verify GPU is available."""ifnottorch.cuda.is_available():
raiseRuntimeError("GPU not available")
@runpod.serverless.register_fitness_checkdefcheck_disk_space():
"""Verify sufficient disk space."""importshutilstat=shutil.disk_usage("/")
free_gb=stat.free/ (1024**3)
iffree_gb<10:
raiseRuntimeError(f"Insufficient disk space: {free_gb:.2f}GB free")
defhandler(job):
job_input=job["input"]
# Your handler code herereturn {"output": "success"}
# Fitness checks run before handler initialization (production only)runpod.serverless.start({"handler": handler})

Key Features:

• Supports both synchronous and asynchronous check functions
• Checks run only once at worker startup (production mode)
• Runs before handler initialization and job processing begins
• Any check failure exits with code 1 (worker marked unhealthy)

See Worker Fitness Checks documentation for more examples and best practices.

Network-Volume Warm Cache

When a network volume is attached, VolumeCache warms local directories (such as a model cache) across cold starts — hydrating them on startup and syncing new files back on exit — so a repeated multi-GB model download becomes a one-time cost per endpoint. It is stdlib-only and best-effort.

fromrunpod.serverlessimportVolumeCachewithVolumeCache(dirs=["/root/.cache/huggingface"]):
model=load_model()

See Network-Volume Warm Cache documentation for configuration and details.

📚 | API Language Library (GraphQL Wrapper)

When interacting with the Runpod API you can use this library to make requests to the API.

importrunpodrunpod.api_key="your_runpod_api_key_found_under_settings"

Endpoints

You can interact with Runpod endpoints via a run or run_sync method.

Basic Usage

endpoint=runpod.Endpoint("ENDPOINT_ID")
run_request=endpoint.run(
{"your_model_input_key": "your_model_input_value"}
)
# Check the status of the endpoint run requestprint(run_request.status())
# Get the output of the endpoint run request, blocking until the endpoint run is complete.print(run_request.output())

endpoint=runpod.Endpoint("ENDPOINT_ID")
run_request=endpoint.run_sync(
{"your_model_input_key": "your_model_input_value"}
)
# Returns the job results if completed within 90 seconds, otherwise, returns the job status.print(run_request )

API Key Management

The SDK supports multiple ways to set API keys:

1. Global API Key (Default)

importrunpod# Set global API keyrunpod.api_key="your_runpod_api_key"# All endpoints will use this key by defaultendpoint=runpod.Endpoint("ENDPOINT_ID")
result=endpoint.run_sync({"input": "data"})

2. Endpoint-Specific API Key

# Create endpoint with its own API keyendpoint=runpod.Endpoint("ENDPOINT_ID", api_key="specific_api_key")
# This endpoint will always use the provided API keyresult=endpoint.run_sync({"input": "data"})

API Key Precedence

The SDK uses this precedence order (highest to lowest):

1. Endpoint instance API key (if provided to Endpoint())
2. Global API key (set via runpod.api_key)

importrunpod# Example showing precedencerunpod.api_key="GLOBAL_KEY"# This endpoint uses GLOBAL_KEYendpoint1=runpod.Endpoint("ENDPOINT_ID")
# This endpoint uses ENDPOINT_KEY (overrides global)endpoint2=runpod.Endpoint("ENDPOINT_ID", api_key="ENDPOINT_KEY")
# All requests from endpoint2 will use ENDPOINT_KEYresult=endpoint2.run_sync({"input": "data"})

Thread-Safe Operations

Each Endpoint instance maintains its own API key, making concurrent operations safe:

importthreadingimportrunpoddefprocess_request(api_key, endpoint_id, input_data):
# Each thread gets its own Endpoint instanceendpoint=runpod.Endpoint(endpoint_id, api_key=api_key)
returnendpoint.run_sync(input_data)
# Safe concurrent usage with different API keysthreads= []
forcustomerincustomers:
t=threading.Thread(
target=process_request,
args=(customer["api_key"], customer["endpoint_id"], customer["input"])
)
threads.append(t)
t.start()

GPU Cloud (Pods)

importrunpodrunpod.api_key="your_runpod_api_key_found_under_settings"# Get all my podspods=runpod.get_pods()
# Get a specific podpod=runpod.get_pod(pod.id)
# Create a pod with GPUpod=runpod.create_pod("test", "runpod/stack", "NVIDIA GeForce RTX 3070")
# Create a pod with CPUpod=runpod.create_pod("test", "runpod/stack", instance_id="cpu3c-2-4")
# Stop the podrunpod.stop_pod(pod.id)
# Resume the podrunpod.resume_pod(pod.id)
# Terminate the podrunpod.terminate_pod(pod.id)

📁 | Directory

.
├── docs # Documentation
├── examples # Examples
├── runpod # Package source code
│ ├── api_wrapper # Language library - API (GraphQL)
│ ├── cli # Command Line Interface Functions
│ ├── endpoint # Language library - Endpoints
│ └── serverless # SDK - Serverless Worker
└── tests # Package tests

🤝 | Community and Contributing

We welcome both pull requests and issues on GitHub. Bug fixes and new features are encouraged, but please read our contributing guide first.