RunCam Feed Camera – Wearable POV & AI Data Capture Camera

The RunCam Feed Camera is a lightweight wearable camera designed for hands-free first-person recording and structured data collection. Capturing 1080P video at up to 60FPS alongside synchronised IMU motion data, it is particularly well suited to AI dataset projects, research, workflow recording, training and documenting real-world tasks from the operator’s point of view. RunCam describes it as an EGO (egocentric) data capture solution designed for repeatable real-world data collection.

Weighing just 26g, the Feed Camera is small enough to be comfortably worn or incorporated into custom mounting arrangements. Depending on the accessories used, it can be mounted to the head, cap or chest, or positioned on a desktop or fixed fixture. Its wide 150° diagonal field of view provides good coverage of the operator’s hands, tools, objects and surrounding workspace while leaving both hands free to perform the task naturally.

This makes the Feed Camera particularly interesting for AI training and egocentric dataset collection. Real-world activities can be captured from the user’s perspective for applications such as human activity datasets, behaviour analysis, data annotation, workflow/SOP recording, motion analysis and research. RunCam’s own example datasets demonstrate tasks including handling objects, writing on paper, using scissors, folding clothing, packing a bag and writing and erasing on a whiteboard.

Alongside the RGB video, the integrated 6-axis ICM-42607 IMU captures gyroscope and accelerometer data at 800Hz together with timestamp information. Raw IMU data can be exported as a separate GCSV file or embedded within the MP4 metadata, providing synchronised visual and motion information for downstream processing.

Importantly, the video and IMU aren’t simply recorded independently. RunCam uses a shared/software-synchronised time base to maintain temporal consistency between the two data streams, specifying a synchronisation error of ≤0.1ms. At 60FPS, approximately 12–15 IMU samples correspond to each video frame. This makes the captured data useful for applications including motion trajectory analysis, electronic video stabilisation and Gyroflow post-processing.

The Feed Camera is equally useful as a straightforward wearable POV recorder. It provides an easy way to record assembly, inspection, maintenance, modelling, DIY projects, demonstrations and other processes without having to hold a camera or repeatedly reposition it. Footage can then be used for review, training, documentation or instructional content.

Video is recorded locally to a microSD card at 1080P 60FPS or 30FPS using H.265 encoding, with cards up to 512GB supported. Normal and loop recording modes are available, while a built-in microphone records audio alongside the video. USB Mass Storage and UVC modes are also supported.

Rather than carrying a large internal battery, the Feed Camera is designed around an external 5V USB-C power source. This helps keep the camera itself exceptionally light while allowing a suitable power bank or the RunCam Power Box to provide extended recording time. RunCam states support for 8+ hours of continuous recording when paired with its Power Box. The small internal battery is intended for power-loss buffering rather than normal standalone operation.

Useful applications for the RunCam Feed Camera could include:

• AI Training & Dataset Collection – Capture first-person RGB video with synchronised IMU motion data.
• Workflow Documentation – Record real-world procedures and complete processes from the operator’s perspective.
• Training & Tutorials – Create hands-free step-by-step demonstrations and instructional content.
• SOP Development & Optimisation – Review how tasks are performed to document or improve standard operating procedures.
• Research & Behaviour Analysis – Collect egocentric video and motion data for research and activity analysis.
• Inspection & Maintenance – Record inspections, servicing, repairs and troubleshooting while keeping both hands free.
• Assembly & Manufacturing – Capture assembly processes, tool use and operator interactions.
• Motion Analysis – Use synchronised gyroscope and accelerometer data alongside video.
• DIY, Making & Crafting – Record electronics, soldering, modelling, crafting and other hands-on projects.
• Remote Review & Technical Support – Capture exactly what the operator sees for later review or troubleshooting.
• Video Stabilisation Research – 800Hz IMU data supports motion analysis and Gyroflow post-processing.
• General POV Recording – A lightweight option for long-duration first-person recording.

Available Versions:

Camera Only / Standard – Includes the RunCam Feed Camera, 1.5m USB-A to USB-C cable, mounting accessories including clip and 1/4″ base, and manual.

Camera Kit – Adds the accessories needed for a convenient wearable setup, including the RunCam Power Box and head strap, alongside the camera, 1.5m USB-A to USB-C cable, mounting accessories, manual and user guide.

Please note: 18650 batteries are not included with the Camera Kit and must be purchased separately. A microSD card is also required and is not included

Resources:

Camera Manual

Camera Kit Manual

Example Data

RunCam provides example recordings demonstrating the Feed Camera in realistic first-person data-collection scenarios. These include handling objects at a desk, writing and cutting paper, folding clothing, packing items and writing on a whiteboard. The supplied examples are recorded at 1080P60 / 18Mbps with 800Hz IMU data, providing a useful reference for the type of RGB + motion datasets the camera can produce.

View RunCam example dataset – handling objects

View RunCam example dataset – writing on a whiteboard