SWE Cloud Agent Workshop

Workshop introduction

What this workshop covers:

  • Running a customizable SWE agent in the cloud, with built in tools: Coding (based on Codex CLI) and Github access
  • Running agents on a managed Backend-as-a-Service platform, designed specifically for AI agents
  • Deploy a Codex-GPT style frontend for managing the SWE agent tasks, logs and status
  • Configure and test your AI Agent through a visual Workbench

End goal: A fully customizable version of ChatGPT-Codex, controlled and configured by you, ready to take on coding tasks, connect to any tool and agent, and be triggered from anywhere.

Solution architecture

Intro to xpander.ai Backend-as-a-Service for agents

Framework-agnostic backend for scalable, real-time AI agents.

FeatureDescription
🛠️ FrameworksUse OpenAI, LangChain, CrewAI, Agno, or raw LLM APIs
⚙️ ToolsPlug-and-play MCP-compatible tools
⚡ EventsBuilt-in streaming: WebUI, Slack, Webhooks
💾 StateMulti-user, distributed out of the box
🚀 HostingAuto-scaling infra, one-command deploy
🛡️ GuardrailsTool dependency safety via agent graphs

Workshop prerequisites

  • On your dev machine, have Python 3.12 & Node.js 22
  • Github account

Ready to build a production-grade SWE agent system? Start with Module 1!