Accident prevention · driver behavior · telematics infrastructure

The accident that doesn't happen

We score phone use, hard braking, and rapid acceleration — then close the loop with coaching that changes what happens next trip. Battle-tested on the budget Android hardware real fleets run. Built by an operator who’s run programs in five markets.

8 yrs In telematics operations
5 markets Programs shipped
1,00,000 km On budget Android hardware

Sastram Labs is a telematics infrastructure company — the SDK that detects driving behavior, the app that shows drivers their own pattern, and the backend that runs the program. We build technology that changes how people drive, not just technology that measures it.

Built by an operator

The program failures that built this company

8 years running telematics programs — usage-based insurance (UBI) pricing, fleet safety, claims — across Colombia, Israel, South Africa, Southeast Asia, and the US. We built the stack we wished we’d had.

Saunak Founder
8 yrs Telematics ops
5 mkts UBI · BBI · Fleet
ex-Zendrive Built the SDK side too
Ituran Group Engineering reference
The architecture

A score without an intervention is just data

Every vendor scores trips. The score goes to a dashboard. The driver never sees it. The behavior never changes.

The score is only valuable if something happens because of it.

Stage 1

Detection

Phone use, hard braking, rapid acceleration — captured on-device, every trip. No driver input. No manual tagging.

Stage 2

Behavioral profile

A single hard brake is noise. A driver who hard-brakes in the top 20%, uses their phone on six trips out of ten, and accelerates aggressively after every stop — that is a risk profile. Built across trips, not just per trip.

Stage 3

Coaching handoff

The profile surfaces in the fleet manager’s dashboard: ranked by driver, broken down by behavior. Not a raw score — something to act on.

Stage 4

Driver nudges

After every trip, Bodha Drive shows the driver their own pattern — phone use, where the hard brakes happened, week over week. A driver who sees their pattern changes it.

Products

The SDK is the product. The rest ships you faster

Most companies start with Sampark and bring their own app and backend. Some take all three because it’s the fastest path from pilot to live program. Take what you need. Replace it later if you outgrow it.

Sampark · the SDK

The core

Native iOS and Android, shared Rust core. Detects trips even after force-kill — the failure mode every other SDK has and won’t tell you about. Battle-tested across the full spread of Android hardware in real fleets, not just flagship phones. The rest is optional.

Read the SDK page iOS · Android · Rust
Bodha Drive · the app

The driver app

Score dashboard, trip history, and post-trip nudges. Use it to launch in weeks, or drop it when your in-house app is ready.

See the driver app Free · iOS & Android
Hastle Free · the backend

The backend

One backend for every partner: drivers, policies, scoring, and notifications out to your own systems. Self-host if data residency requires it. Connects Sampark or any other SDK.

So you’re never locked in
How we work

A telematics SDK is not a product

A live UBI or BBI policy needs four things to work. The SDK is one of them. We do all four — because we’ve been the customer of an SDK that did only one, and watched the program stall.

01

Use case & jurisdiction

UBI, BBI, fleet safety, claims — each implies a different scoring model and consent flow. We start with the regulator’s spec.

02

SDK + app integration

Permission flows, workarounds for phone makers that aggressively kill background apps, and recovery when they do anyway — the part nobody writes about, which is the part that breaks programs in production.

03

Calibration

Thresholds calibrated against your pilot fleet — your devices, your geography, your drivers. Not a synthetic test set.

04

Launch operations

The first 90 days of a program are operational, not technical — and we run them with you.

Honest envelope

Software is imperfect.
We’ll show you exactly where.

Most telematics pitches lead with one accuracy number. We don’t — because that number is meaningless without your devices, your geography, your drivers. Here’s what we’ll actually tell you on the first call.

Trip detection 1,00,000km

Tested where SDKs go to die

Logged on the OEMs whose process killers break most SDKs in production. We name where any SDK degrades — and recalibrate against your pilot before a policy goes live.

Budget Android OEMs Aggressive process killers Real fleet hardware
Battery & sensor cost 0% idle CPU

The honest envelope

We tune for low impact; heavy commuters see more. We’ll show you the exact envelope on the first call — not the number that looked best on a slide.

Motion sensor off at idle GPS coarse No idle drain
Time to launch 30min · first call

Pilot-driven, not pitch-driven

We commit a date once we understand your stack — and we put the engineer who’ll do the work on the call that commits it.

Engineer-led Pilot-driven
Integration

Initialise once. It runs

One entry point. Configure it with your tenant key and trips start flowing within the first drive. If your team has integrated CMT or Zendrive before, you’ll be productive in an afternoon.

SamparkClient.swift
// One initialisation. Same shape on iOS and Android.
import Sampark

let sampark = Sampark.configure(
    tenant: "acme-insurance",
    region: .in_south,
    onTripStart: { trip in
        analytics.log("trip.start", trip.id)
    },
    onTripEnd: { trip in
        backend.submit(trip)
    }
)

// On user login.
sampark.setUser(id: currentDriver.id)
sampark.start()

// That's it. Trips detect, score, upload — even after force-kill.
I’ve been on both sides of this — running telematics programs as the operator, and building the SDKs they ran on. Most telematics tools are built by engineers who’ve never had a fleet client call them at 11 pm because trips are missing. We built Sampark having run those programs first, in five markets, for insurers and fleet operators with real money on the line.

That’s not a longer pedigree. It’s a different one.
Saunak · Founder, Sastram Labs 8 years in telematics · ex-Zendrive Connect on LinkedIn
Engineering proof
“Sastram took over backend infrastructure end-to-end. The engineering judgment was a tier we don’t see often.”
Ituran Group on a separate engagement, where Sastram took over their backend infrastructure end-to-end
FAQ

Questions, answered

Do I need all three products?
No. Sampark is the core — most companies integrate it into their own app and backend. Bodha Drive and Hastle Free exist for teams that want to move faster without building an app or backend from scratch.
What Android devices has this been tested on?
The budget Android phones real fleets actually run, not just flagship models — 1,00,000 km logged across four tiers of aggressive, battery-killing hardware, plus a stock Android and iOS reference for comparison. See the device matrix for the breakdown.
Do you sell driver data?
No, and we don’t use it to serve advertising. Detection and scoring run on-device by default — see our Privacy Policy and Data Residency pages for the full detail.
What happens on the first call?
30 minutes with the engineer who’d own your integration. Bring your device list, geography, and timeline — we’ll tell you exactly what to expect on your fleet, or refer you to someone better suited.
Talk to an engineer

Tell us what you’re trying to launch

30 minutes with the engineer who’d own your integration. We’ll tell you exactly what to expect on your fleet — or refer you to someone better suited. No follow-up sequence.

Book a call

Engineer on the call · Your devices · Your geography · Your timeline