openma-ai/open-managed-agents — Open-source Claude Managed Agents API implementation and self-hosted Claude Tag-style agent runtime.

Description and use

**Brief Summary (2‑3 sentences)**
openma‑ai’s **open‑managed‑agents** is an open‑source implementation of Claude’s Managed Agents API that lets you run Claude‑style “Tag” agents either on Cloudflare Workers/Durable Objects or in a Node.js environment. It provides a drop‑in compatible runtime for orchestrating multi‑agent, tool‑using workflows, complete with API/SDK/CLI interfaces and built‑in memory handling. Licensed under Apache 2.0, the project is actively maintained and already used in several pilot deployments.

**Value**

• **From prompts to repeatable workflows** – Turns isolated Claude prompts and tool calls into deterministic, version‑controlled agent pipelines, reducing drift and manual glue code.
• **Unified orchestration** – Enables multi‑agent coordination, tool‑use pipelines, and standardized memory stores without reinventing the underlying runtime.
• **Flexibility of deployment** – Works on edge (Cloudflare Workers/Durable Objects) for low‑latency use cases or on traditional Node.js servers for more heavyweight processing.

**Practical Adoption Path**

1. **Prototype** – Clone the repo, run the provided CLI or SDK locally to wrap an existing Claude prompt as a Managed Agent.
2. **Integrate** – Replace ad‑hoc API calls in your application with the Open‑Managed‑Agents SDK, defining agents, tools, and memory schemas in TypeScript.
3. **Deploy** – Choose the environment that matches your latency/cost profile:

• *Edge*: Deploy the Worker/Durable Object bundle to Cloudflare for ultra‑fast, globally distributed execution.
• *Server*: Deploy the Node.js package to your existing backend or container platform.

1. **Scale & Observe** – Use the built‑in logging and health endpoints, add monitoring (e.g., Cloudflare Analytics or Prometheus), and iterate on agent definitions.

**Production Readiness**

• **Activity & Community** – 184 ★, 16 forks, recent commits (as of 2026‑06‑26), and a healthy set of 20 topics indicate an active codebase.
• **Maturity** – The API/SDK/CLI surface is stable, with clear TypeScript typings and documented deployment patterns for both edge and server runtimes.
• **Risk Profile** – No major metadata or licensing concerns (Apache 2.0); the remaining checks are standard security and maintainer responsiveness reviews.

Overall, the project is considered **highly ready for a serious pilot** and can be promoted to production once the organization completes its usual security and compliance vetting.

Problem

openma-ai/open-managed-agents helps turn isolated prompts and tools into repeatable agent workflows.

Outcome

High for an OSS candidate: recent activity, adoption, and ecosystem signals are strong enough for a serious pilot.

Workflow

• coordinate multi-agent workflows
• add tool-use pipelines
• standardize agent memory

Evidence

Open-source Claude Managed Agents API implementation and self-hosted Claude Tag-style agent runtime. Drop-in compatible; runs on Cloudflare Workers/Durable Objects or Node.js. Apache 2.0.

Source