Case studies/Adlytic AI
In-store audience analytics from cameras already on the floor
Computer vision for retail: who is in the aisle, how long they stay, which displays get attention — without installing a new hardware stack.
- Client
- Adlytic AI
- Location
- United States
- Year
- 2025
AI · SaaS
info@lyrocode.com
Input
Existing CCTV
Output
Store metrics
Year
2025
Pictures from the work
Photography that matches the product — the buildings, the floor, the clothes, the road — not stock of people pointing at a laptop.
The job
Retail teams already have cameras. They did not want a new box on every ceiling. They wanted counts, dwell time and display attention from footage they already collect — with a dashboard a store manager can read, not a research notebook.
The model has to work across lighting, camera angles and store layouts. Privacy matters: no faces in the product the client sees.
What we built
We built the vision pipeline to sit on existing RTSP and file drops. People and zones, not identities. Store-level dashboards in a React app: today’s floor, this week versus last, which endcap is dead.
Training and evaluation stay in Python. The product the retailer logs into is a SaaS console — sites, cameras, reports — not a Jupyter export.
Where it landed
Retailers get aisle and display numbers from cameras they already paid for. The console is something a regional manager can open on Monday. Faces never leave the processing boundary the client agreed.
- Vision pipeline on existing store cameras
- Zone, dwell and display metrics
- Retailer dashboard and site admin
- Privacy-preserving processing path
Stack
What we shipped with — picked for this product, not for a slide.
Python
PyTorch
FastAPI
React
Next.js
PostgreSQL
AWS
Docker
“We were tired of vendors who wanted us to rip out the cameras. Lyro Code used what was on the wall, and the dashboard is something my ops people actually open.”
More work
Willis Bond·Web · Property
Project sites a development team can actually keep current
Marketing and project pages for a New Zealand developer — precinct photography, availability, and content the marketing team can change without a ticket.
Read the story
Maria B.·Commerce · Web
A fashion storefront that has to survive sale season
Commerce work for Maria B. — lookbooks, collections, and a checkout path that still holds when a drop hits.
Read the storyTell us what you are building.
We usually reply within one business day. Email info@lyrocode.com if you would rather skip the form.
