Pipeline Integration

Bring automated CPET interpretation into the data pipeline

Oxynet is designed to process CPET time-series programmatically and return structured interpretation outputs. The next step is to move automated interpretation closer to where CPET data are acquired, processed, and used.

From exported files to continuous data flows.

No manual CSV or XLS transfer required.

A test is recorded, and the interpretation follows, with nobody deciding one file at a time to run it.

Current State

Today: interpretation often starts after export

File-based analysis is effective for research, validation, and individual testing. At scale, however, export and manual transfer create an unnecessary boundary between CPET acquisition and automated interpretation.

CPET Device

Exercise laboratory

Manufacturer Software

Data acquisition & display

Test Completed

CPET time-series recorded

Export & manual transfer

CSV / XLS Export

File-based data handoff

Manual File Handling

Local storage, email, shared drives

External Platform / Upload

Data re-entry or file upload

Oxynet

Automated interpretation layer

Structured Outputs

Thresholds, intensity domains, metrics

Oxynet currently enters the workflow after data have been exported and manually handled. That is necessary today, but not architecturally required.

The Goal

Interpretation as part of the pipeline

Existing acquisition and data-management workflows can remain unchanged. Oxynet can operate as an interpretation layer, receiving CPET time-series programmatically and returning structured outputs for downstream systems.

CPET Device

Exercise laboratory

Manufacturer Software / Core Lab

Acquisition & data infrastructure

CPET Data Pipeline

Existing infrastructure

Existing workflow

Continues unchanged

↓ as before

Oxynet layer

Oxynet API

Automated interpretation layer

Automated Interpretation

VT1, VT2, intensity domains

Structured Outputs

Programmatic response

Clinical / Research / Software

Downstream systems

Designed to integrate, not replace.

Oxynet does not replace acquisition systems, manufacturer software, or existing clinical workflows. It adds an automated interpretation layer that operates on the same CPET time-series.

No manual CSV or XLS transfer required.

When integrated at the data-pipeline level, CPET time-series reach Oxynet programmatically, without any intermediate export or manual file handling.

How It Works

Three steps, one integration point

01

CPET time-series in

Breath-by-breath or other supported CPET time-series are transferred programmatically from an existing data system to the Oxynet API.

02

Oxynet interpretation

Oxynet parses the vendor file as exported: twenty formats, each with its own unit conventions and clock quirks, then measures the physiology in it.

03

Structured outputs out

Thresholds, per-breath intensity domains, derived quantities, oscillation analysis and signal-integrity checks are returned as structured JSON for review, storage, visualisation or downstream analysis.

Oxynet is reachable over REST, over MCP for AI assistants, or via the Python package for local research pipelines. The API documentation lists every endpoint and the twenty vendor formats detected automatically. Oxynet's outputs are not intended to replace clinical review. They provide a consistent, reproducible measurement baseline.

Collaboration

Do you operate a CPET data pipeline?

We are interested in working with organisations that acquire, process, or manage CPET data at scale, including exercise testing core laboratories, clinical trial networks, hospitals, research infrastructures, software providers, and CPET technology partners.

Core Laboratories

Exercise testing centres managing continuous or high-volume CPET data flows.

Clinical Trial Networks

Multicentre studies requiring consistent, automated CPET interpretation across sites.

Software & Device Partners

CPET software providers and device manufacturers exploring interpretation integration.

Research Infrastructures

Academic and clinical networks evaluating automated interpretation in real-world workflows.

We are particularly interested in prospective integration, continuous data flows, multicentre validation, and evaluating automated interpretation within real-world CPET workflows.