# Need more coding agents? just spawn them on the cloud w/ Oz by Warp Source: https://www.youtube.com/watch?v=0HotPgUCzyo Recap page: https://rapidrecap.app/video/0HotPgUCzyo Generated: 2026-02-27T23:03:40.457+00:00 --- ## Quick Overview The video demonstrates how to use Oz by Warp to manage and run coding agents in the cloud, enabling developers to automate tasks like running scheduled news aggregators, setting up development environments, and handling PR fixes remotely, effectively increasing productivity by shifting compute load off local machines. **Key Points:** - Oz by Warp allows users to run AI coding agents in the cloud, which can be triggered via CLI, web UI, scheduled cron jobs, or integrations like Slack/Linear. - The platform was used to set up an AI News Curator project, which required creating a dedicated Warp environment using a specific Docker image (warpdotdev/dev-base:1). - Scheduled agents, like the daily news check, run autonomously in the cloud, ensuring work continues even when the developer closes their laptop (0:05). - The Oz dashboard allows users to monitor all agent runs, including scheduled, interactive, and team-based tasks, showing credit usage and run times (2:13, 5:22). - The demonstration included refactoring a repository, setting up a recurring schedule for the news aggregator (every 5 minutes initially, then updated to every 5th day), and authorizing GitHub access for commit/PR pushing (3:37, 6:50, 7:26). ![Screenshot at 2:15: The Oz dashboard showing multiple running or completed agent tasks, including 'Refactor Scout Repo Utilities' and 'Create AI News Aggregator Project,' demonstrating the centralized monitoring capability of the platform.](https://ss.rapidrecap.app/screens/0HotPgUCzyo/00-02-15.jpg) **Context:** This video serves as a tutorial and showcase for 'Oz by Warp,' an orchestration platform designed to run cloud-based AI coding agents. The narrator contrasts the traditional developer workflow, which often involves manual 9-to-5 coding and being tethered to a local machine, with the efficiency gains offered by Oz. The core concept revolves around using agents to handle repetitive tasks, environmental setup, and background processes, freeing up the developer for higher-level work. ## Detailed Analysis The video advocates for using Oz by Warp to leverage cloud-based AI agents for software development tasks, arguing that this dramatically increases efficiency compared to relying solely on local machine coding. The narrator sets up an AI News Curator project as a case study, demonstrating how Oz handles environmental setup, dependency installation, code refactoring, and creating scheduled tasks. Key features highlighted include the ability to spawn agents on demand or schedule them using cron syntax (e.g., cleaning up stale feature flags every four days). The platform provides a unified control plane via the Warp app (CLI, web, or mobile) to start, track, and monitor all agent activity, including team collaboration on runs. The environmental setup process is fully automated, analyzing the repository structure (e.g., Next.js/TypeScript project) to select the appropriate base image and determine necessary setup commands like 'pip install -r requirements.txt'. The scheduler allows agents to run autonomously in the cloud, even when the developer is offline, which is shown by scheduling the news aggregator to run every 5 minutes initially, then updating it to run every 5th day. The platform tracks resource usage via 'credits' and provides a dashboard to view all agent runs, confirming that complex multi-step tasks (like repository setup and code modification) are completed successfully in the cloud. ### Introduction to AI Agents - Comparison of AI coding efficiency (exponential growth) vs. manual coding (diminishing returns) - AI agents are becoming essential for tech giants (0:39, 0:50) ### Oz Agent Overview - Oz is an orchestration platform for cloud agents that are programmable, auditable, and steerable - Agents can run on the cloud, turning skills into automations (1:32, 2:39) ### Trigger Mechanisms - Triggers include Oz scheduler (cron), tagging in Slack/Linear, manual start in Warp, API/SDK, or CLI (3:12, 3:20) ### Agent Infrastructure - Agents run on Oz Infrastructure (hosted runners) or User's Infrastructure (self-hosted runners) (3:15) ### AI News Curator Project Setup - The agent was used to initialize a Git repo, analyze the Python project structure, determine setup commands ('cd ai-news-curator && pip install -r requirements.txt'), and create a cloud environment using Docker (4:46, 5:54) ### Scheduling Agents - Demonstrated creating a recurring schedule using 'oz schedule create' for daily checks, and subsequently updating it to run every 5 minutes using cron syntax, all viewable and manageable via the Oz dashboard (6:50, 7:26) ### Agent Capabilities & Results - Agent successfully refactored code, fixed Cloudflare blocking issues by removing xAI, and aggregated news headlines from sources like Anthropic and Google DeepMind into the README file (6:38, 7:50) ![Screenshot at 0:01: Two cloud-headed characters sitting on beanbag chairs illustrating a conversation about coding workflows.](https://ss.rapidrecap.app/screens/0HotPgUCzyo/00-00-01.jpg) ![Screenshot at 0:13: A stylized hacker figure overlaid with terminal code and the Google logo, representing high-level tech work being accelerated by AI.](https://ss.rapidrecap.app/screens/0HotPgUCzyo/00-00-13.jpg) ![Screenshot at 0:35: A diagram illustrating that Tech Giants are built upon AI/LLMs \(represented by the OpenAI logo\), implying that not using AI leaves one behind.](https://ss.rapidrecap.app/screens/0HotPgUCzyo/00-00-35.jpg) ![Screenshot at 1:14: A split-screen view showing a terminal session interacting with an AI agent \(left\) and the resulting code/environment setup \(right\).](https://ss.rapidrecap.app/screens/0HotPgUCzyo/00-01-14.jpg) ![Screenshot at 4:34: Warp pricing tiers \('Build' at $18/month providing 4 CPUs/4GB RAM, and 'Business' at $45/user/month providing 4 CPUs/8GB RAM\).](https://ss.rapidrecap.app/screens/0HotPgUCzyo/00-04-34.jpg)