Plano is an AI-native proxy server and data plane for agentic applications.
Building agentic applications for production requires infrastructure beyond the initial prototype: routing requests to the correct agent, enforcing safety guardrails, capturing observability signals, and managing model provider differences. Plano addresses this by centralizing these concerns into a unified out-of-process dataplane rather than scattering them across application code and frameworks. The tool provides agent orchestration with low-latency routing, model agility through name-based or preference-driven routing, zero-code capture of agentic signals with OpenTelemetry traces and metrics, and moderation and memory hooks via filter chains for safety policies. Built on Envoy by its core contributors, Plano decouples applications from framework abstractions and allows teams to use any language or AI framework.
Plano suits teams building production agentic systems who want to avoid reimplementing routing logic, observability instrumentation, and safety mechanisms across codebases. It works best for organizations that need to add new agents without modifying application code and want consistent guardrail and moderation policies across their agent fleet. The tool is particularly valuable when model agility matters, as it centralizes routing decisions rather than embedding them in agent implementations.
Almost all open issues are raised by outside users rather than the core team, indicating a substantial base of adopters reporting real-world use. Maintainers typically respond to new issues and pull requests within a day. Work in the issue tracker is dominated by enhancement requests and bug reports, with routing functionality receiving particular attention.
