iTrace

iTrace is an open-source Python toolkit for analyzing gaze and pupil data, with optional webcam capture and local viewing tools. It best suits researchers and developers exploring eye-movement methods; its analysis breadth is useful, but its webcam measurements are not validated against real eye-tracking devices.

Overview

At its core, iTrace processes gaze CSV files with NumPy- and SciPy-based analysis, calculating gaze in degrees of visual angle and identifying eye-movement events. Webcam capture is an optional layer, so the analysis workflow can also be used with existing gaze data. That separation makes it a flexible reference implementation, rather than a complete substitute for validated tracking hardware.

The project reports algorithmic checks against synthetic ground truth, but no validation against real eye-tracking data or a reference tracker. Researchers making device-level accuracy claims should compare results with a reference device or public dataset; synthetic checks alone do not establish real-world performance.

Key features

Gaze events and visualizations

For fixations and saccades, iTrace offers I-VT and I-DT detection; microsaccades use the Engbert–Kliegl method. Views include gaze traces, x/y time series, speed plots, event rasters, scanpaths, direction polar plots, amplitude histograms, and main-sequence scatter. The range supports inspection from several angles, though interpretation still depends on the quality and validation of the input data.

Pupil analysis and calibration

Pupil tools cover blink detection and interpolation, preprocessing, and eye-crop segmentation. Segmentation reports pixel size or pupil size relative to the iris, not diameter in millimetres. The latter is a relative proxy, so it can support algorithm exploration but should not be treated as a calibrated physical measurement. Affine screen-space calibration reports calibration error and gaze-quality measures, useful diagnostics that do not resolve the project's real-device validation gap.

Capture, interfaces, and exports

Optional webcam capture uses OpenCV and MediaPipe Face Mesh iris landmarks. A local FastAPI and WebSocket HTML app and a Streamlit dashboard provide viewing options; the live app can export derived gaze, pupil, and capture CSVs alongside a JSON analysis report. The CLI can analyze gaze CSVs and export JSON reports, or record gaze and pupil CSV data from a camera.

The default workflow does not save raw eye video or persisted eye-crop images, which limits retained sensitive imagery. Live viewing may send a transient JPEG eye crop to the browser over a local WebSocket, so users should account for that data flow. Python installation is documented through uv, with capture, dashboard, figures, and web components available as optional extras.

Pricing

iTrace (MIT-licensed open-source software) costs nothing and is an open-source Python package under the MIT license. There are no paid tiers in this plan structure. It is a straightforward fit for users comfortable installing Python software and assembling the optional components they need; those seeking a validated commercial tracker should look elsewhere.

Platforms

iTrace supports Linux and self-hosted use, with a web browser interface served locally. The browser option is tied to the local HTML interface rather than a separate hosted service. Webcam is the listed input method, while CSV analysis also supports gaze data already collected elsewhere.

Who it's for

Researchers and developers who want an inspectable reference implementation of gaze-event and pupil-analysis methods are the strongest fit. It is also relevant to people prototyping webcam workflows or analyzing gaze CSVs without requiring capture. It is not the right choice when the requirement is demonstrated accuracy against real eye-tracking hardware, or pupil diameter in millimetres.

Pros and cons

• Pros: Multiple established event-detection methods and a broad set of visualizations support detailed algorithm exploration.
• Pros: Capture is optional, and the CLI can work with gaze CSVs, so camera use is not required for analysis.
• Pros: Default exports omit raw video and persisted eye-crop images, reducing retained image data.
• Cons: Synthetic and closed-loop checks do not establish performance against real devices.
• Cons: Webcam pupil size is only a relative proxy, not a millimetre measurement.
• Cons: Optional capture and interface components add dependencies to an otherwise focused scientific analysis core.

Alternatives

For a broader set of eye-tracking options, see Eye Tracking Software.

• GazePlotter is another free, open-source option, with support across Android, iOS, Linux, macOS, web, and Windows.
• GazeRecorder is a free alternative for Android, iOS, web, or Windows users; its free plan is limited to non-commercial use.
• Labvanced may suit researchers who want a broader web-accessible platform; its free plan includes 300 MB storage, 10 recordings, and one active recruiting link.
• Hawkeye is an option for iOS or web use with a small free allowance of one project and five sessions; its Starter Plan costs $39.99 a month.
• PyGaze is another free, open-source Python toolbox, for users targeting Linux, macOS, or Windows.
• Chiasm is a paid web and API alternative with a free trial and usage-based Flex pricing, for those who prefer a commercial service model.
• OGAMA is a free, open-source alternative for Windows users.
• OpenGazer is another free, open-source gaze tracker for Linux and self-hosted use.

Verdict

Choose iTrace if you need an open-source, inspectable toolkit for experimenting with gaze and pupil-analysis methods or analyzing gaze CSVs. Its strongest case is the range of analysis methods and optional capture and interfaces; its decisive limitation is the absence of real-device validation. For accuracy claims or physical pupil measurements, use a reference-validated system instead.