About Midstreamly
Houston-based. Four-person team. Angel-backed. We build condition monitoring for midstream rotating equipment — because the signal is already in your historian.
Beatriz Herrera spent 11 years as a reliability engineer at a Houston-area midstream pipeline operator. In 2022, a centrifugal compressor on a gas gathering line failed — an unplanned shutdown that cost the operator 38 hours of throughput and a six-figure repair bill.
After the investigation, she pulled the OSIsoft PI historian data. The bearing degradation pattern was there — visible in retrospect — for six days before the machine tripped. Spectral amplitude at the bearing defect frequency had been climbing. The vibration trend was unmistakable. There was no tooling in place to watch for it in real time.
She left operations to build that tooling. Midstreamly was incorporated in 2024 with one core premise: the data already exists in your systems. We connect to it, analyze it, and alert your reliability team before the machine fails.
"Every midstream operator I've worked with has a PI historian. Almost none of them have a team watching the vibration data in real time. That gap is what we exist to close."
These aren't marketing copy. They're the product decisions we make every week.
We connect to what you already have — AVEVA PI, Bently Nevada, Emerson AMS — via read-only API. No new sensors. No PLC changes. No six-month hardware procurement cycle. If your historian is running, we can start within weeks.
Read-only integrationWe publish our false-positive rate (<0.3% on anomaly alerts). We don't hide it in fine print. If our detection accuracy degrades on your equipment population, we tell you — and we work to fix the model, not explain away the failures.
Published FP rate · Model transparencyWe align with PHMSA integrity management program documentation and API 670 machinery protection conventions — not because we're required to, but because midstream equipment failures aren't just operational problems. They're safety events.
PHMSA IMP · API 670 alignmentFounder & CEO
11 years as a reliability engineer at a Houston-area midstream pipeline operator. API 670 certified. Expertise in centrifugal compressor and pump condition monitoring. Founded Midstreamly in 2024.
ML Engineer
MS Data Science, UT Austin. Specializes in time-series anomaly detection and survival modeling. Leads the Midstreamly ML pipeline from OSIsoft PI tag ingestion through LSTM and Weibull model deployment.
Reliability Domain Expert
14 years at gas processing plants across the Permian Basin and Eagle Ford. CMRP certified. Leads Midstreamly's failure mode library — documenting spectral signatures for centrifugal compressors, multistage pumps, and gas turbine drivers.
Field Deployment Engineer
IT/OT integration specialist. Deployed data collection infrastructure across air-gapped compressor stations throughout Texas, Louisiana, and Oklahoma. Expert in OPC-UA bridging and OT network segmentation.
If you're a reliability engineer or plant manager dealing with unplanned compressor or pump downtime, we'd like to hear about your setup — even before you're ready to evaluate a vendor.
Start the Conversation