Oz

The platform includes a comprehensive shared data plane that stores and shares agent context across sessions, agents, and harnesses. This includes conversation transcripts, artifacts, skills, rules, and the innovative Agent Memory system. Agent Memory represents a fully portable, cross-harness memory system that helps Claude Code, Codex, and Warp agents remember what works across every session, preventing repetition of the same mistakes. This shared persistent memory allows agents to automatically draw on real team experiences to become more efficient and effective, essentially creating self-improving agents that turn feedback and institutional knowledge into better skills, prompts, tools, and evaluations.

Oz operates through a sophisticated technical approach that combines a unified control plane, agent execution layer, and data plane for running and managing agents with built-in shared context, permissions, observability, orchestration, and memory. The platform supports multiple integration points, allowing users to interact with cloud agents through Warp Terminal, Web, or Mobile interfaces, and trigger them from various sources including Slack, Linear, GitHub, or directly via CLI, API, or SDK. Agents can run in sandboxed environments hosted by Warp or on your own infrastructure, providing flexibility for different security and compliance requirements.

The benefits for users include measurable outcomes such as increased automation efficiency, reduced manual intervention, improved agent coordination, and enhanced visibility into automated processes. Teams can scale their automation efforts from simple scheduled tasks to complex multi-agent workflows without worrying about infrastructure management or context loss. The platform's enterprise-grade observability features enable tracking of every agent run in real time with comprehensive steerability and tracing built in, giving teams confidence in their automated systems and the ability to intervene when necessary.

Concrete use cases demonstrate Oz's practical applications, such as when a Slack-triggered agent kicks off a cloud run and developers can steer the session from Warp Terminal to polish changes before they land. Another example involves complex long-horizon tasks like large feature builds where multiple agents work in parallel on different components, with a supervisor agent coordinating their efforts. Codebase migrations represent another strong use case, where Codex agents can systematically transform code while Warp Agent orchestrates the overall migration strategy and handles exceptions or complex transformations.

Target users include developers and engineering teams at companies of all sizes, particularly those already using Warp Terminal and needing to scale their automation efforts. The platform integrates with popular development tools and platforms including Slack, Linear, GitHub, and supports enterprise requirements like SOC 2 Type II compliance, Zero Data Retention policies, SSO, and SCIM. While specific pricing plans aren't detailed in the provided content, the platform offers both cloud-hosted and self-managed deployment options, with agents able to execute on machines you provision inside your cloud, VPC, or data center for maximum control.

In summary, Oz represents a comprehensive solution for organizations seeking to leverage AI agents for software development automation at scale. By providing the necessary infrastructure for coordination, context management, and observability, it enables teams to move beyond isolated agent interactions to create truly collaborative agent workforces. The platform's emphasis on seamless transitions between local and cloud execution, combined with its sophisticated orchestration capabilities and persistent memory system, makes it a powerful tool for modern engineering teams looking to automate complex workflows and improve their development processes through intelligent agent coordination.

Updated 2026-02-28