← sources

Building Master Agent

source: reddit (reddit:AI_Agents)published 2026-10-26 21:44ingested 2026-10-27 02:35

retained text

Looking for advice from people who build AI agents in code



I'm planning a "super agent": one master agent that manages a team of agents, one lead agent per business, each with its own specialists. Before we start building, I want to get the foundation and the hosting right.



What matters to us:

\- We own everything (code, data, hosting). No lock-in to one platform.

\- The AI model is swappable (Claude, GPT, Gemini, local models).

\- Built-in human approval before anything risky (money, contracts, public posts).

\- Long-running jobs that can pause and resume.

\- Strong security. We ruled out one platform after a serious vulnerability this year.



Right now LangGraph is the front-runner, with Microsoft Agent Framework as the backup.



Questions:

1. Is LangGraph the best option for this in your experience, or would you pick something else?

2. Where would you host it? Your own cloud (AWS / GCP / Azure), LangGraph Platform, or something else?

3. Anything you wish you'd known before building multi-agent systems in production?



Thanks, all real-world experience welcome.

document fetched 2026-10-27 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1xga6hz/building_master_agent