LekoTech delivers fast search on a global scale

THEIR CHALLENGE

Digitizing a complex, analog industry

For the founders of LekoTech, an automotive startup initially launched in Albania, the mission was simple yet ambitious: to digitize the traditionally highly analog ecosystem of spare parts.

The goal was to make every vehicle owner's life easier by creating the industry’s largest dataset of car parts. More importantly, LekoTech aimed to advance the circular economy by focusing exclusively on original equipment manufacturer (OEM) parts, which are manufactured to last a lifetime.

“LekoTech’s primary aim is to help breakers and dismantlers manage parts and inventories digitally,” said Nikolin Ngjela, CTO and Co-founder at LekoTech. “Our Fleet Management app is a consumer-oriented layer that lets vehicle owners and fleet operators access that data for everyday maintenance and parts sourcing.”

By encouraging dismantlers to “reduce, reuse, and recycle,” the company aims to significantly cut global carbon emissions and reduce the environmental damage caused by counterfeit and discarded parts. However, aggregating resale inventory in this sector is highly complex: a single spare part can fit anywhere between 10 and 50 specific vehicle models, creating huge amounts of metadata that must be instantly searchable against hundreds of car configurations, marketplace listings, and warehouse locations.

The industry is also highly traditional; typical customers can be third-generation, family-owned breakers’ yards with millions of parts but relatively little technological know-how.

“Dismantlers often don’t have a close connection with technology,” explained Ngjela. “One of our biggest struggles as a startup was to both educate them and make our system easy to use.”

To make this vast, fragmented catalog accessible to buyers and fleet operators, speed is paramount. LekoTech previously struggled with search latencies of up to 35 seconds, a delay that risked losing potential buyers to competitors if they couldn’t instantly confirm a part’s price and availability.

Furthermore, because users search for parts using highly variable terminology—ranging from exact OEM codes to natural language descriptions—LekoTech needed a robust infrastructure capable of handling high levels of data variability, while scaling rapidly across Europe.

“The smallest details in components often make the biggest difference,” said Ngjela. “And you often don’t see the differences until a mechanic removes them. That’s the level of ambiguity that we need to manage.”

OUR SOLUTION

A unified data platform that delivers

Recognizing that its complex, unstructured datasets would not suit a rigid relational database, LekoTech chose MongoDB as its core operational foundation.

“The data we work with doesn’t really fit into tables. A spare part document needs to include vehicle metadata and applicability across hundreds of car configurations,” said Ngjela. “MongoDB is very scalable, and it’s a perfect fit for our solution. That’s why we’ve used it from day one.”

And as the company expanded from Albania into eight Italian regions, it also recognized how MongoDB’s technical architecture can support long-term growth and high-volume marketplace integrations.

To optimize its search capabilities, LekoTech worked closely with MongoDB Professional Services to continuously refine its data structure and performance efficiencies. Through this collaboration, LekoTech is now transitioning away from dedicated vector databases to consolidate its search strategy onto MongoDB Vector Search on Atlas.

By unifying its architecture on MongoDB, the company is also able to simplify its engineering experience. “Before that, we had to manage 32 or more standard indexes in the database,” added Ngjela. “And it was delivered in slow queues.”

With the transition to MongoDB Vector Search now 75% complete, LekoTech utilizes a single Atlas search index. In addition to significantly accelerating responses, the resulting unified platform has paved the way for advanced generative AI integrations, with LekoTech using MongoDB Vector Search alongside Voyage AI to power semantic, natural-language search capabilities. Instead of relying solely on exact OEM part numbers, customers can now simply type a query such as ‘door 2007 BMW’, and the system instantly understands the context and retrieves the correct part. It’s a critical differentiator for LekoTech’s users.

“The system understands that type of prompt, and responds as you need it to,” added Ngjela. “Whether you’re the vehicle owner, or a dismantler or warehouse operator, it transforms the search experience.”

OUTCOME

Lightning-fast searches drive reuse, resale, and growth

With MongoDB Vector Search on Atlas, LekoTech is driving both its operational performance and commercial trajectory. The business has reduced its search latency from 35 seconds to under two seconds, and in many cases, full inventory searches across 300,000 spare parts can be handled in less than 100 milliseconds.

And by providing traditional dismantlers with an intuitive, operating system-like platform, LekoTech enables smaller businesses to export inventories automatically to the world’s biggest e-commerce platforms with a single click. The results have been staggering.

“We have real case examples where customers have doubled their sales within two or three months,” said Ngjela. “And they see the potential for even more, especially now that with MongoDB, they can apply AI tools to their business.”

These unprecedented levels of speed and accuracy are also driving growth for LekoTech. Its robust technical foundation on MongoDB proved critical in passing rigorous evaluations to become an official eBay partner in Europe, a milestone that helped the company achieve €2 million in eBay sales in a single year.

LekoTech’s focus on the circular economy and environmental sustainability has also opened doors to strategic partnerships with major OEM across Europe.

“We’ve been presenting the solution to big manufacturers, and it’s been getting lots of interest,” added Ngjela. “Spare parts are manufactured to last a lifetime, and you can cut carbon emissions significantly by reusing them. OEMs increasingly understand that from an environmental perspective, this is very, very important.”

Operating a fully compliant, scalable, and highly stable architecture also allows LekoTech’s engineering team to focus on innovation and expansion. Already active across Northern Italy and expanding into Germany, Spain, and France, the company is now eyeing the massive US market.

“MongoDB has given us the flexibility we needed as a startup to develop the solution and the ability to grow at scale,” said Ngjela. “It’s very effective and trustworthy, and that makes all the difference for us.”