Working with renewable energy clients needed a modern orchestration platform that could handle the complexity of distributed BESS site infrastructure. The industry was largely fragmented across SaaS point solutions, but the client needed something they could standardize on. A single platform capable of ingesting real-time IoT sensor data streamed through VPN tunnels from active BESS sites directly into ClickHouse. Fred Rivollier the Founder and CTO at Librenergy chose Mage, deploying it as the foundation for the client’s entire data stack.

With Mage in place, three distinct workflows now run in production.

• The core ETL pipeline continuously pulls high-frequency IoT readings from BESS sites, writing performance data by the minute into ClickHouse for near real-time monitoring.

• A second workflow automates the analytics layer, running DBT jobs, executing notebooks, and delivering scheduled reports via email without any manual intervention.

• A third handles DevOps automation, using Mage as a reliable orchestration layer for infrastructure backups and operations that previously required separate tooling.

The most advanced workflow shows what modern AI-native pipelines can look like in practice.

Using Claude Code agents inside Mage, the team built an autonomous battery performance engineering pipeline where a Mage block runs Claude as a subprocess. Claude connects to a ClickHouse MCP server, queries daily battery fault data, and identifies anomalies like voltage irregularities, abnormal charge cycles, and cells trending toward failure. It then produces a complete diagnostic report covering site-level performance summaries, flagged battery units, likely fault causes, and recommended next steps, delivered automatically to stakeholders each day. One simple prompt, full production output.

Learn more about Librenergy