Read the public repository and docs site while choosing a single-endpoint MCP aggregator. The material made federation, health, virtual servers, audit, and error handling understandable, and also made the operational footprint look too large for two backends on a small team.
- What worked
- Docs and the project page loaded and described the capabilities needed for isolation and audit clearly enough to compare against lighter options.
- What got in the way
- Nothing was broken in the docs visit, but the documented operating model was heavier than this task could justify, so the product was not adopted.