How We Built an Autonomous Coding Agent on Astropods
What if your coding agent wasn’t trapped on your laptop, locked to a single model provider, or cut off from the rest of your engineering team? In this episode, Senior Developer Advocate Pooja Mistry sits down with Rodric Rabbah, Head of Product and Co-Creator of Astropods, to break down Ada, a cloud native autonomous coding agent built directly on the Astropods platform. Ada is a great example of a custom AI agent deployed in production, with a secure sandboxed terminal in the cloud, freedom from model lock in through the Astropods AI Gateway (swap between Claude, Kimi K3, and self hosted models), and “shared intelligence” that compounds across your whole development team. What You’ll Learn in This Episode: -The limits of local only coding agents: tethered laptops, single provider lock in, and unshared context -Meet Ada, a cloud sandboxed autonomous coding agent named after Ada Lovelace -Two ways to interface with Ada: a custom terminal frontend (AG-UI) or the built in Astropods Chat playground -Why shared intelligence beats isolated local workflows (the central heating and plumbing analogy) -How Ada was used to build and iterate on Ada itself -Model routing with the Astropods AI Gateway: Anthropic, open weights, and on prem options -How to deploy Ada from the Blueprint Registry with automated GitHub CI/CD syncing -Access controls, identity management, and persistent memory with PostgreSQL Resources & Community: Try Astropods and explore the Blueprint Registry: https://astropods.com Join the community on Discord: https://astropods.com/discord #Astropods #AIAgents #CodingAgent #AutonomousAgents #CustomAgents #DeployAIAgents #ClaudeCode #DevTools #SoftwareEngineering #OpenSource #DevOps #llmops 0:00 - Introduction: Meet Rodric Rabbah & Ada the Coding Agent 0:35 - Why Move Coding Agents to the Cloud? 1:25 - Naming Ada & The Philosophy of Agents as Teammates 1:53 - Live Demo: Inspecting Ada's Cloud Sandbox & Terminal UI 3:06 - Built-In Astro Chat vs. Custom Web Frontends (AG-UI) 4:35 - "We Used Ada to Build Ada": Self-Improving Agents 5:05 - Breaking Free from Model Lock-In (Astro AI Gateway & Open Weights) 5:57 - Shared Intelligence: The Central Plumbing Analogy 7:40 - Enterprise Cost & Work Tracking 8:35 - Deploying Ada from the Blueprint Registry & GitHub Syncing 9:50 - Access Controls, Memory DBs & Hardware Sizing 11:34 - Launching Ada & Final Thoughts