Healthier streets
Fewer piles means fewer mosquitoes, rats and stray animals, and less smoke from burning.
AI garbage detection for India's towns
Every street. Every pile.Tracked till it's cleared.
Our AI device pins every garbage pile with photo, location and time, shares it with contractors and verifies the clean‑up.
Book a demoGarbage · High
8th Main × 11th Cross
Malleshwaram, Bengaluru© OpenStreetMap contributors
Every town has that corner.
It starts with one bag.
Left overnight. Nobody reports it, because everybody assumes someone else will.
It grows.
Garbage attracts more garbage. One bag becomes a dumping spot.
It spreads.
Dogs scatter it, rain carries it into drains, and drains flood in the monsoon.
It's burned.
When it gets too big, someone sets it on fire, and the whole street breathes it.
Now picture every street in India.
0.0 lakh
tonnes a day
of garbage across India
0%
collected
Collection has come a long way
0%
unaccounted for
Neither treated nor landfilled
Source: CPCB Annual Report on Solid Waste Management, 2022–23. Illustrative: dots show the scale of the problem, not measured locations.
No two piles are the same.
Swachathe has been trained to know what garbage on a street looks like, every kind of it, from a moving vehicle.
Garbage · one pile, many kinds
01 · What kind of waste
- Wet
- Dry
- Sanitary
- Special care
- Rubble
02 · What it's made of
- Plastic
- Food
- Paper
- Glass
- E-waste
- Cloth
03 · Where it is
- On the road
- At a wall
- In a drain
- By a full bin
04 · How bad it is
- Low
- Medium
- High
- Area in m²
So we built
A device that never stops looking for garbage.
Meet swachathe.
One device. Four parts working together.
- Dedicated camera
- AI computer
- IoT sensors
- GPS & storage
Dedicated camera
Films the roadside continuously as the vehicle moves.
Wide-angle capture sees the whole kerb, and motion-blur correction keeps every frame sharp, even at speed and in harsh light.
AI computer
Watches the video right there on the vehicle and spots garbage.
Detections happen in real time, on the device. Nothing waits to be uploaded and checked later.
IoT sensors
Keep the device powered and connected.
A fast data link sends detections the moment they're made, and battery management keeps the kit running all shift.
GPS & storage
Records exactly where every pile is.
Every issue is logged by its address on the map, road to road, across thousands of images.
From video to the city's screen, in seconds.
All of this runs on the device, in real time, while the vehicle is still moving. The city sees every vehicle, live.
On the device · on the move
✕ Blurry frame dropped
● RecordingArea · 3.2 m²
Severity · HighGPS ✓ · Time ✓
- CaptureThe camera films the street, non-stop.
- FramesThe device pulls out still frames and drops the blurry ones.
- DetectAI draws a box around every pile of garbage it sees.
- SegmentIt traces the pile's outline to measure how big it is.
- Score & tagIt scores how bad it is and stamps the place and time.
- SendThe record reaches the city's platform while we're still driving.
City dashboard · Live4 vehicles on route
© OpenStreetMap contributors● New detectionGarbage · from V1Photo, GPS and time attached
● New detectionGarbage · from V1Photo, GPS and time attachedEvery vehicle carrying our device, live on the city's map: where it is and every pile it finds.
From the street to the city's desk.
Step 01 / 05 · Survey
The same ward, street by street.
A camera on a moving vehicle covers the same ward, street by street, on a regular cycle.
- One camera
- Same route
- Same criteria
Step 02 / 05 · Detect
Garbage, found in real time.
It has been trained to know what garbage on a street looks like. Faces and number plates are blurred on the device.
- AI on the device
- Every frame checked
- Privacy blur
Step 03 / 05 · Tag
A map pin, a time stamp and a picture.
Each pile is tagged with its picture, location, time, severity, outline and a unique ID.
- Picture
- Location
- Time
- Severity
- Segmentation
- Unique ID
Step 04 / 05 · Report
Straight to the city.
Detections reach the city's platform while the vehicle is still driving. The official assigns each one to a contractor, with nearby jobs bundled into one trip.
- Live ward map
- Assigned to a contractor
- One trip, many jobs
Step 05 / 05 · Verify
Nothing is marked done without proof.
The contractor uploads a photo of the same spot as proof of work, and the next survey checks it again.
- Proof photo
- Rechecked next survey
- Repeat spots flagged
Garbage · High
Rechecked · clear
- Class
- Garbage
- ID
- #574
- Confidence
- 0.53
- Severity
- High
- Location
- 13.0049° N, 77.5703° E
- Time
- 11 Oct, 10:42 AM
Swachathe platform · your ward
New pile · HighAssignAssigned
Proof photo
✓Proof of work uploaded
Spotted a pile we missed?
Residents can report it from the citizen web app. It lands on the same platform, with the same photo, location and tracking.
Open the citizen web app ↗This is what our camera sees.
Boxed, outlined and scored as the vehicle passes. Real captures from our device.
Our process finishes. Today's keeps ticking.
The same pile, two ways. Watch the day counter.
Day0
With Swachathe
In progressDone ✓- A pile appearsDay 0
- Detected, tagged, on the city's screenDay 1
- Assigned to a contractorDay 1.5
- Cleared, proof photo uploadedDay 3
- Verified on the next surveyDay 4
✓ Proof of work✓ Status at every step
Today
Waiting…Still open- A pile appearsDay 0
- Someone finally complainsDay 2
- The complaint reaches the wardDay 3
- Cleared, if it's picked upDay 9
✕ No proof✕ No status for the citizen
Approximate days, for illustration. To be confirmed with pilot data. Assigning and clearing are done by the city and its contractors; Swachathe makes every step visible.
Why towns choose Swachathe.
Accurate
The same detection standard is applied to every street we pass.
Consistent
Each ward is swept on the same route, to the same criteria.
Reliable
Every issue arrives with a photograph, a coordinate and a time.
Scalable
Adding a ward means adding a device, not more manual inspection.
swachathe
Real-time monitoring
Detections reach the platform while the vehicle is still driving.
Faster response
Nothing waits for a complaint before it reaches a contractor.
Cost savings
Jobs are bundled into one optimised trip instead of many.
Accountable
Open issues and delays are visible ward by ward.
A clean street is where a growing town starts.
Safer monsoons
Garbage that never reaches the drain can't block it when the rains come.
Money well spent
Jobs are bundled into one optimised trip instead of many.
A town to be proud of
Clean towns rank higher in Swachh Survekshan and give pourakarmikas the credit they deserve.
Built for Swachh Bharat. Aligned with the UN's goals for cities.
What India and the UN ask of every town, and how Swachathe helps meet it.
Garbage Free Cities
At least 3-star for every town
Every open pile found and tracked, so garbage-vulnerable points get closed.
Swachh Bharat Mission–Urban 2.0
Star Rating
Proof for assessors
A geo-tagged, time-stamped photo and status for every pile.
GFC Star Rating Protocol, 2022
Swachh Survekshan
A better rank every year
Month-on-month improvement a town can point to.
MoHUA annual survey
SWM Rules 2026
Every stage tracked online
A digital record of every pile found and cleared.
In force 1 April 2026
UN SDG 11 & 12
Measure uncollected waste
Street-level evidence for the hardest part of indicator 11.6.1.
UN Sustainable Development Goals
Starting in Karnataka.
- Bengaluru WestMalleshwaram wardPilot starting soon
- ArsikereTown municipal councilGoing live soon
KarnatakaArabian SeaBengaluru WestArsikereYour town next?
- Launched at Bengaluru Tech Summit, Government of Karnataka, 2024–25
- Grassroots innovation award, Karnataka [confirm exact award name]
Our camera looks for garbage. Not for people.
Camera view
Face blurred on the device
Garbage
Blurred before it's saved.
Faces and number plates are blurred on the device, in real time, in line with the Digital Personal Data Protection Act, 2023. Unblurred faces and plates are never stored.
Only what's needed.
We keep the photo of the pile, where and when it was seen, and how it was cleared. Nothing about the people who pass by.
Locked when offline.
If the network drops, records wait in an encrypted store on the device and sync when it returns.
Let's make your town the next clean one.
For municipalities
See Swachathe on your own wards. We'll show you the device, the platform and what it finds.
For residents
Spotted a pile on your street? Report it from the citizen web app and follow it until it's cleared.
Open the citizen web app ↗Is the camera watching me?
No. It looks for garbage. Faces and number plates are blurred on the device before anything is saved.
Who clears the pile?
Your city's own teams and contractors. Swachathe finds each pile, reports it to the city and keeps track of it until it's gone.
Does my town need new equipment or staff?
No. Our device does the surveying; your teams use the Swachathe platform to assign and close each pile.
How fast will a pile be cleared?
That's up to the city's teams. Swachathe makes sure they know straight away and shows how long each pile stays open.
Does it work at night?
Not yet. It works in daylight for now, and the camera handles glare, shade and wet roads.
Which towns have Swachathe?
A pilot is starting soon in Malleshwaram ward, Bengaluru West, and Arsikere is going live soon. Want your town next? Book a demo.