Introducing Gitpod SDK:
you focus on agent development,
we handle your infrastructure

Try now Learn more

We provide enterprise-ready infrastructure, you bring the agents.
Gitpod's SDK takes your SWE agents from experimental to deployable at scale.

Example SDK Usage (Python) 
import asyncio
from gitpod import AsyncGitpod
import gitpod.lib as util

repo_url = "https://github.com/containerd/containerd"
command_to_execute = "go test -v -short ./..."

async def execute_command_in_environment():
    client = AsyncGitpod()

    machine_class_id = (await util.find_most_used_environment_class(client)).id

    create_params = {
        "machine": {"class": machine_class_id},
        "content": {"initializer": {"specs": [{"context_url": {"url": repo_url}}]}}
    }

    environment = (await client.environments.create(spec=create_params)).environment

    await util.wait_for_environment_running(client, environment.id)

    async for line in await util.run_command(client, environment.id, command_to_execute):
        print(line)

    await client.environments.delete(environment_id=environment.id)

if __name__ == "__main__":
    asyncio.run(execute_command_in_environment())

What can you build with the SDK?

Unlike running AI assistants locally where they can do anything on your machine, Gitpod development environments are completely isolated and come with everything needed to work with code: your preferred editor, complete access to git history, build tools, and secure connections to your internal systems. Get the freedom of local AI development with the security of cloud sandboxes. Currently only available for Gitpod Flex users.

Agents draft the PR,
you finish in a real editor

Trigger the launch of AI agents from your GitHub issue, platform CLI or Backstage. When the agent is done, pick up in Cursor or VS Code and add any final touches. It's the dev experience of v0 but handed off to an editor you actually enjoy using.

More than 'looks good to me' PR reviews

Schedule your agents to run overnight and wake up to fully validated, production-ready PRs. Replace basic 'comment only' automated reviews with agents that understand your code. Instead of only scanning for issues, your agents can run tests to validate and fix changes.

Dependabot is annoying, configure an LLM instead

Go beyond simple version bumps to true continual dependency maintenance. Replace basic dependency scanner PRs with intelligent agents that don't just flag updates - they understand your codebase, update dependencies, fix breaking changes, and ensure tests pass before you even look at the PR.

Discover Gitpod SDK. Deploy agents at scale in your infrastructure now.

Learn more

You should not build this yourself

Building your own infrastructure to orchestrate agents is complex, and both Kubernetes and Docker are the wrong abstractions

01

Secure container runtime

Isolate processes to manage namespace, and constrain container privilege to enable Docker-in-Docker development

02

Editor integration

Manage complex IDE connectivity, handle file system event, and maintain editor state synchronization

03

State management

Control full environment lifecycle, handle persistence layer, and maintain reliable performance

04

Storage orchestrator

Control disk allocation, disk IO bandwidth and IOP management, and handle all backup/restore operation

05

Image distribution system

Optimize container image operation and caching for reduced startup time and efficient storage

06

Resource scheduler

Allocate CPU, memory, and network resource with fair distribution algorithm and latency optimization

07

Network security manager

Isolate network environment, manage port forwarding, and enforce granular access control

08

Process supervisor

Monitor running process, handle failure, and maintain system stability across environment

09

Integration and authentication

Control access and integrate with source control system, secret management, and handle machine-to-machine authentication flow

10

Metrics collector

Track detailed performance metrics, resource utilization, and comprehensive cost data

11

Environment builder

Create and update workspace image within defined security boundary and compliance control

12

Logging and monitoring

Capture agent execution flow, log streaming, and visibility into agent behaviour and decision-making processes

What's next?

While you’ll need to build solutions using our API/SDK, we’ve eliminated the months of infrastructure work standing between you and production-ready agents. Imagine promoting Devin from junior engineer to staff?

Discover Gitpod SDK.

Deploy SWE agents at scale.

Learn more