A single binary that replaces your entire camera evaluation toolchain — live streaming, real-time analysis, OpenCV pipeline, hardware diagnostics. Plug in a V4L2 camera. Run ./videre. Evaluate. Used by individual engineers and hardware teams shipping on embedded Linux.
two-thread capture/processing15 pixel formats · H.264 · Bayerx86 · Raspberry Pi 5 · Jetson Orinno Qt · no Electron · no display serverREST + WebSocket API
In Action
See it. Don't just read about it.
ROI Analytics
Draw a rectangle. Get per-channel stats — live.
Select any region on the live frame. Per-channel min/max (R, G, B, luma) and horizontal/vertical line profiles compute and update at camera frame rate — no script, no capture step, no post-processing.
per-channel min/maxline profileslive stream
Image Analysis
Live Histogram
R/G/B/luma distribution rendered on every captured frame.
Temporal Analysis
5-Second Image Quality Report
Collect 5 s of frames. Get an automatic report: over/underexposure, colour temperature, noise, and flicker frequency via DFT.
OpenCV Pipeline
Processing — No Code Required
Blur, sharpen, threshold, grayscale, Canny edges — toggle from the sidebar. Runs on every frame in the processing thread.
Capabilities
Everything the job needs. Nothing it doesn't.
Videre replaces the fragmented toolchain — guvcview, throwaway Python scripts, the Windows vendor utility — in one binary that runs on your production hardware.
Deployment
Single Binary, Zero Runtime Dependencies
No Qt, no Electron, no desktop environment. Web UI embedded at compile time. Runs headless on dev boards and over SSH tunnels.
./videreheadlessSSHPWA
Image Analysis
Live Histogram + Temporal Stability
R/G/B/luma distribution live on every frame. Analyze 5 seconds of data: exposure report, colour cast, flicker frequency via DFT, clipping indicators.
histogramflicker DFTROI analysis
OpenCV Pipeline
Filters, Detectors, Barcode/QR — No Code
Gaussian, Canny, binary threshold, sharpen. Contour, circle, blob detectors run in parallel across CPU cores. ZBar-based QR/barcode decoder live on every frame.
edge detectblob detectQR/barcode
Pixel Formats
H.264 Decode, Bayer Demosaic, 15 Formats
MJPEG, H.264 via FFmpeg, YUYV, NV12, GREY, Y16, RGB24, BGR24, and four 8-bit Bayer mosaic patterns demosaiced to RGB24. If V4L2 outputs it, Videre shows it.
H.264Bayer×4Y16MJPEG
Hardware Diagnostics
USB Speed, Bandwidth, VA-API — with Fix Hints
Dedicated diagnostics page reads sysfs without root. USB bus speed, autosuspend state, bandwidth utilisation, VA-API availability. Actionable recommendations, not raw numbers.
USB busbandwidth %VA-API
Multi-Platform
x86, Raspberry Pi 5, and Jetson Orin — Same Codebase
Runs natively on x86, Raspberry Pi 5, and NVIDIA Jetson Orin (Nano, NX, AGX). Full ISP support on RPi 5 — Camera Module v1–v3 and HQ Camera (IMX477) tested. On Jetson, dedicated Argus ISP controls (AE, AWB, denoise, edge enhance, frame timing) and Max Performance mode for native-resolution preview at the highest achievable frame rate.
x86Raspberry Pi 5Jetson OrinArgus ISPIMX219 · IMX477 · IMX708
Before & After
Stop cobbling together a toolchain.
The gap between "camera plugged in" and "confident go/no-go decision" shouldn't take hours of scripting, tool-switching, and driver archaeology.
Before Videre
✕v4l2-ctlCryptic ioctl output. Manual format enumeration. No GUI, no histogram, no analysis.
✕guvcviewShows a picture. Segfaults. No controls for half your cameras. No temporal analysis.
✕Python + OpenCV scriptTwo hours to write. Covers one camera. Lives in /tmp. Re-written every evaluation.
✕Windows vendor utilityRequires a separate machine. Results don't match production Linux hardware.
✕sysfs inspectionUSB speed and bandwidth dug out manually, raw numbers with no interpretation.
With Videre
✓All V4L2 controls, enumerated on openEvery driver-reported control surfaced as sliders, menus, and toggles. Camera-reported ranges.
✓Live histogram + temporal stabilityR/G/B/luma live. Flicker detected via DFT. Exposure and clipping report in 5 seconds.
✓OpenCV pipeline, no code requiredCanny, blob, barcode/QR — toggle from the control panel. Parallel across CPU cores.
✓Runs on your production hardwarex86, Raspberry Pi 5, and NVIDIA Jetson Orin supported. Camera Modules v1–v3 and HQ Camera tested with full ISP pipeline. Argus ISP controls on Jetson.
Whether you're running a camera evaluation team, shipping developer kits, or integrating Videre into your bring-up workflow — there's a tier for that.
Team License
Up to 10 or 25 engineers.
For in-house vision teams doing repeated camera evaluation, integration work, and customer support — without the per-seat friction of individual licenses.
Bundle Videre on dev kits, co-market to your customers, and earn affiliate commission on every referred sale. Designed for hardware vendors who want a better first-hour experience out of the box.
✓ Bundle Videre Community on dev kit SD images
✓ Co-marketing: case study, webinar, partner listing
Make a bare sensor behave like a machine vision camera.
The app shows you the frame. The SDK builds the camera. libvidere turns a bare MIPI sensor module into a machine vision camera in software, and a CUDA ISP gives Jetson clean, measured image quality without Argus lock-in. Validated on Jetson Orin Nano and Raspberry Pi 5. In early access, not yet generally available.
Camera Runtime
libvidere: a bare sensor, as a capable camera
Trigger, multi-camera sync, per-frame metadata, stable identity, and tuning, provided in software when the sensor lacks them and routed to PiSP, Argus, or NVENC when the platform has them. One API across Raspberry Pi 5, Jetson Orin, and generic V4L2.
C / C++ / PythonSFNC featuresRPi5 + Jetson
GPU Image Quality
Videre CUDA ISP for Jetson, no Argus lock-in
Develops raw Bayer and mono to clean RGB and NV12 on the GPU, with portable libcamera and DCP tuning, noise-model denoise, and image quality checked in CI. 4K at 30 fps with zero-copy input on an Orin Nano, with the CPU left free for inference.
4K @ 30 fpszero-copy inputΔE / SSIM gated
Portable ISP API
One ISP control model, any backend
The same create, configure, process, and control calls over CUDA, PiSP, Argus, and a CPU fallback, with runtime stage introspection. The ISP is the part of a camera stack that breaks hardest on every port, so this is the seam that survives a change of sensor or platform.
CUDA / PiSP / Argus / CPUbackend-neutralintrospection
Design Partner Program
Early access · Jetson Orin + Raspberry Pi 5
Building on a bare sensor and feeling the bring-up and tuning pain? We are taking one or two design partners: camera module vendors and vision system OEMs with a real sensor on Jetson or Pi. You get the SDK on your own hardware and a direct line to the developer while the API is still soft enough to shape.
Headless servers, SSH sessions, automation pipelines. Every feature in the UI is also available from the command line or REST API.
bash
# List connected cameras $ videre list /dev/video0 Logitech C920 uvcvideo usb-0000:00:14.0-1 /dev/video2 ZWO ASI120MM uvcvideo usb-0000:00:14.0-2
# Capture a lossless TIFF, skip 30 frames for auto-exposure to settle $ videre capture/dev/video0--format tiff --skip 30 --output shot.tiff
# Record 30s on a remote Jetson, stream to local file $ sshuser@jetson"videre record /dev/video0 --duration 30 --output - "> clip.avi
# Headless server + REST API control $ videre--no-browser --device /dev/video0 & $ curllocalhost:8080/api/stream/start-X POST $ curllocalhost:8080/api/save-X POST -d '{"format":"tiff"}'
Pricing
The software costs less than the camera.
Download Community and evaluate a camera in the next five minutes. No account, no license key, no setup.
Single Camera
Community
Free
Personal use & evaluation
✓ All features enabled
✓ Latest release always
✕ Commercial use
✕ Usage data collection cannot be disabled
✕ Requires internet connection
Check your inbox!
Anonymous usage data collection is mandatory
Pro Most Popular
€89
One-time · perpetual · single user
✓ All features
✓ Commercial use licensed
✓ No usage data collection
✓ No internet required
✕ Updates (snapshot at purchase)
Pro includes 3 lifetime device activations. Reinstall on the same machine will usually reuse the existing activation, but major system changes may consume another credit. Support can help in exceptional cases.
Subscription includes 5 device activations per year, renewed annually. Reinstall on the same machine will usually reuse the existing activation, but major system changes may consume another credit. Support can help in exceptional cases.
