The signal platform, built as one integrated stack.
One resolver. One source of truth. One API.
How it works
Pipeline notices and scheduled-quantity flows enter through the same ingestion layer, get standardized through the same resolver, and emerge through the same open API. A change at the pipeline meter level can move a downstream signal in seconds.
For the trading desk, this means the platform plugs into the workflow that already exists. The desk doesn't replace what's already there; it adds a signal layer that watches the physical reality the existing stack can't see. Severity-graded events surface where the trader already works — in the dashboard, in the alert stream, eventually in the Excel model or the Snowflake query.
For the risk and audit teams, every interpretation in the platform is reproducible. The original notice is one click away from the resolved event. The standardization is deterministic — events run against versioned rules and content-hashed processes that reproduce the same output from the same inputs. A trader, a risk officer, and an auditor can all stand in the same record of what happened, when it happened, and why it mattered.
How it's built
PipeProwler is the physical-flow surveillance product at the base of the stack. Pipeline notices and scheduled-quantity flows enter through PipeProwler's ingestion, get resolved into severity-graded events with full lineage, and become the substrate every downstream capability depends on. The PipeProwler page covers the product in depth.
Under the hood, the platform is a single integrated data stack. Raw ingestion lands in a bronze layer. Structured extraction and resolution produce a silver layer. Aggregated, query-ready gold tables and materialized views drive the dashboards and the API.
Topology, hub tags, flow direction, and capacity are governed as data, not hardcoded in application logic; the resolver is the single source of truth for what every meter means, and topology can be updated without redeploying the platform. Delivery is API-first — the same data surface drives the dashboard, the programmatic integration, and any future partner distribution.
The architecture is built to accommodate additional capabilities — market-intelligence and synthesis layers — through the same ingestion, resolver, and API surface.
The flagship
PipeProwler is the foundational product in the stack — the place where the physical reality of the U.S. natural gas grid becomes a signal the platform's downstream capabilities can consume.
It surveils the pipelines and corridors that move Henry Hub prices, including every active Gulf Coast LNG export corridor, in real time. Pipeline notice text and scheduled-quantity flows are ingested continuously, standardized into a unified topology of meters and hub tags, and resolved into severity-graded events with full data lineage from raw source to API.
Every downstream capability the platform supports draws from PipeProwler's resolved topology. The platform is one stack with PipeProwler at the base, not a catalog of separate products.
Delivery
Delivery is API-first by design. The same data surface that drives the platform's dashboard drives both the programmatic integrations the buyer's existing stack consumes through and the alerting channels that surface events to the desk in the moment.
Programmatic, today
REST API with documented endpoints across the full data surface: physical-flow events, the underlying topology, and the full lineage records.
More, on roadmap
Python SDK for quant and data-science workflows. Excel Add-in for analyst and trader desks that operate in spreadsheet-native environments. Snowflake share for institutions that maintain their own data warehouses and want the platform's data resident alongside the rest of their stack.
Advanced alerting
Custom alert routing over email, SMS, Slack, Microsoft Teams, and webhooks. Alert routing is configured at onboarding to the buyer's specific severity thresholds, corridors of interest, and notification destinations. Additional channels are available on request.
Methodology and trust
Methodology
The platform's data sources, standardization, quality controls, and the severity rubric itself are visible to the buyer through customer documentation, and the lineage trail every event carries. Severity scoring is deterministic: every event runs against versioned rules and content-hashed processes that reproduce the same output from the same inputs.
Trust
Coverage figures and event exhibits are measured against production, not estimated. Every event carries the operator's posting time, our first-seen time, and our processing time, so latency claims can be verified against our own data. Thirteen independent monitoring layers watch the surveillance itself, and every operator index is reconciled against our capture. SSO/SAML available for institutional deployments. SOC 2 in progress; data lineage and audit logs available on request. No customer data sold or shared with third parties, and no customer data used to train external models.
Talk to the team.
ThermSuite is sold through direct engagement with natural gas market participants. Reach out for a working demo of the platform — the integrated stack, the resolved topology, and the API surface shown live.