Video telematics for service providers
Don't just record the road. Understand it.
Peregrine adds contextual visual intelligence to your existing cameras and products, enabling better safety, risk and claims services on your platform.
Because context matters
Events are noise. Context is insight
Peregrine reads the relationships between road, vehicle and driver to reveal what is happening, why it matters and how urgently to respond.
The risk score is calculated on the device, frame by frame, and travels with the event into your platform.
Peregrine risk score0 = routine · 100 = act now
* Illustrative example, used to show how the reading changes as signals are added. Full notes.
Detection is just the beginning...
Recognise how road, vehicle and driver interact, from close following and cut-ins to distraction, vulnerable road users and low-visibility hazards.
Both moments were risky. Only one was the driver's doing.
A G-force trigger logs both brakes the same way. Peregrine keeps two readings apart: how dangerous the moment was, and how much of it the driver caused.
14:02
Braked hard behind a cyclist
Reduced visibility, a wet surface and a vulnerable road user entering the lane. A dangerous moment, and the driver read it early.
The conditions created the risk. Do not penalise the driver.
* Illustrative example. The risk score describes how dangerous the moment was; driver attribution describes how much of it the driver caused. Full notes.
Both readings fuse
Seeing it run is quicker than reading about it. We will walk you through real trips, scored the way they would be in your platform.
Book a demoWhat contextual understanding changes
Fewer false alerts
Drivers stop ignoring the system, so coaching means something.
Less data moved
Your cellular bill stops scaling with every camera you ship.
Lower claims costs
Footage and context that show what actually happened.
Fewer accidents
Risk signalled while there is still time to avoid it.
Turn better understanding into better products.
For fleet operators
Insight into how drivers adapt to changing road conditions, rather than a list of times the accelerometer moved.
For drivers
Good driving recognised, and resilience in bad conditions rewarded instead of scored as risk.
For insurers
Trip scoring with the context attached, so a claim can be answered with why it happened and not only how.
See it. Understand it. Act on it, inside your product
Peregrine delivers a structured, anonymised event into your existing backend, user experience and customer workflows.
Why it triggered
Conditions read on the device
The same event as the motorway brake above, as your customer would see it.
We provide
- Structured, scored events
- Video and metadata, anonymised before storage
- Real-time and historical access
You build
- Longitudinal tracking
- Driver and fleet profiles
- Analytics and dashboards, under your brand
Your integration team can read the endpoints today. Full documentation is public, no account needed.
Read the API documentationFlexible deployment
Upgrade the intelligence. Keep the hardware.
Deploy on supported smartphones, dashcams and embedded camera systems without forcing customers into a new hardware or platform strategy.
Ready to deploy
Available nowMultiple viewpoints
Multi-camera systems
Front, side, rear and cabin views resolved into one reading of the situation, so broader coverage does not mean more video for somebody to watch.
- Best fitTrucks, buses and complex fleet operations
- ConnectsUp to five additional cameras on a single device
- DeliversPrioritised events with evidence from the view that matters
- DevicesMiTAC K265 · DVR and edge gateway setups
Road and driver
Connected dual-facing cameras
One device watching the road and the cabin, so an event carries both halves of the story into the telematics product your customers use.
- Best fitVans, delivery fleets and commercial vehicles
- UnderstandsRoad context, driver state, and how the two relate
- DeliversIn-cab signalling and structured events over the API
- DevicesMiTAC K265 · Peregrine One · MiTAC K220 · your own device on request
- We recommendThe K265 above also runs dual-facing, and it is the one we would put forward: better camera, more compute, more headroom.
Mobile and compact devices
Road intelligence on smartphones and lightweight Android cameras. The quickest way to run a pilot, and the route for passenger vehicles and cost-sensitive programmes.
Android Automotive
Android Automotive runs in the car itself, so there is no aftermarket device to fit. We reach the vehicle through the operating system rather than through a direct integration with the manufacturer.
Already have a camera or compute platform in mind?
Tell us the chipset, operating system and camera configuration, and we will map the most direct route to deployment.
Your product stays yours.
Integrate contextual events, on-device models or the complete perception stack while retaining control of your brand, platform and customer relationship.
| Included | Plain EdgeOn-camera software | Edge + APIPlus hosted events | TurnkeyPlus connectivity |
|---|---|---|---|
| Road-facing ADAS | |||
| Driver monitoring | |||
| Live streaming | |||
| Feature configuration | |||
| Cloud video processing | |||
| Video provision over API | |||
| Anonymisation before storage | |||
| Global connectivity | |||
| Mobile data plan |
Plain EdgeYou run your own data pipeline and want the perception only.
Edge + APIYou want API-ready events without building storage and retrieval.
TurnkeyYou want to launch without setting up connectivity at all.
On Plain Edge you own the pipeline, so anonymisation and storage sit with you. Cloud and mobile data consumption beyond an agreed volume affects pricing.
Not sure which tier fits? Tell us how much of the pipeline you run today and we will point you at the right one.
Book a demoWill it run on your camera?
Tell us the device and we will send you our hardware requirements, so your engineers have something concrete to check against. If you would rather we looked at it, one of ours will come back on your specific device.
What we look at
- Whether the processor has headroom to run perception continuously
- Camera resolution, frame rate and field of view
- How updates reach the device once it is in the field
Hear from the teams building on Peregrine
Telematics, insurance, automotive engineering and research partners explain what Peregrine makes possible in real products and services.
Bridgestone · September 2025"This is a major step in our mission to make mobility safer and smarter."
Jan-Maarten de Vries, President, Fleet Management Solutions, Bridgestone
Webfleet Video ships Peregrine perception as a paid upgrade, retrofitted to commercial vehicles already in service over the air.
Read the announcement
Linqo GmbH · May 2025"Our customers want systems that just work—and work together."
Max Donders, Managing Director, Linqo
Peregrine Vision runs on the dual-camera dashcams Linqo distributes, with alerts landing in the dashboard fleet managers already use.
Read the announcement
Allianz · published case study"With the help of Peregrine, we are able to solve these problems."
Stefan Sellschopp, Allianz Partners, Munich
Allianz approached Peregrine because autonomous driving was too new to price with confidence. The work was about reading traffic situations, detecting how they change, and judging the risk that follows.
Read the case study
Porsche EngineeringPorsche Engineering runs Peregrine object detection on a smartphone inside its ComBox app to check what the vehicle's own traffic sign recognition reports. Where the two disagree, the phone photographs the sign and tells the data logger to capture the vehicle data. Porsche says the setup was tested during development of the new Macan.
Read it on Porsche EngineeringPut it on one of your cameras
Send us a device and a use case. We will tell you what runs on it, what does not, and what a pilot would look like.
Questions from integration teams
What hardware does it run on?
Connected dashcams, Android Automotive devices and smartphones. The models are processor-agnostic within minimum requirements for compute, memory and camera specification, so for a specific device we need the specification first.
What happens when the vehicle loses connectivity?
Perception runs on the device, so detection, scoring and in-cab signalling carry on with no connection at all. Events and clips queue locally and sync when the vehicle is back in coverage.
What does the API return?
Structured events with trigger reasons, event metadata, video clips, device status and location, and GPS traces for completed trips. Webhooks push notifications as events occur. Endpoint documentation is at api.peregrine.ai/docs.
Who handles personal data?
On Edge + API and Turnkey, faces and number plates are removed before footage is stored, on EU servers. On Plain Edge you run the pipeline, so anonymisation and retention sit with you. Under 1% of frames leave the vehicle in any case.
Can we ship it under our own brand?
Yes. We sell to telematics providers and platforms rather than to fleets. The software is delivered as an SDK your team integrates, events arrive in your platform under your name, and your customers deal only with you.
How long does an integration take?
For a supported device with an existing pipeline, weeks rather than months.* Timelines depend on your platform and on how ready your integration team is, so we scope it before quoting one.
Notes
- * Accident and claims figures were measured by Peregrine in live customer deployments. Method and sample are described in the Peregrine.ai whitepaper.
- False alert reduction and data reduction were observed in selected deployments and vary with camera placement, route profile and configuration.
- Risk scores, driver attribution and the trip example are illustrative. The risk score describes how dangerous the moment was; driver attribution describes how much of it the driver caused. They are separate readings.
- Integration timelines depend on the platform and on integration readiness.