They say what they need.Navlu finds the product.
A shopping assistant that knows your catalog. Navlu understands what your customer needs and finds products that fit their budget and the conversation.
Your catalog. Your style. One snippet.
In the example store, a visitor opens the shopping assistant, types a gift request and receives two recommendations. They view the first product, add it to the cart and see a confirmation.
The right product. Now in their cart.
From a message to a match.
Relevant products come first. Jev evaluates the options against what your customer needs.
A crowded collection of products appears as the customer types “A gift for my partner under 4,000 TL. No plush toys.” Irrelevant products disappear as the sentence is completed. Pressing Enter arranges the remaining candidates, then Jev evaluates them and three suitable products are selected.
SETUP
From catalog to conversation.
Connect your products
JSON, CSV, XML or PostgreSQL. Keep prices, stock and availability up to date with scheduled feeds.
Make it yours
Customize colors, product cards, mascots and the requests your assistant can handle.
Go live on your store
Add the embed snippet. Let customers discover products through conversation, on mobile and desktop.
PAY AS YOU GO
No subscription.
You're in control.
Add credit and track usage in your dashboard.
for about 4,000 product searches.
Search volume varies with conversation context and the number of products evaluated.Create an accountFAQ
Short answers.
- What is Navlu?
- Navlu is a conversational product discovery assistant for ecommerce stores. It first finds relevant candidates from your catalog, then uses TypeSafe's Jev model to evaluate and rank them by the customer's intent, budget and conversation context.
- How is Navlu different from sending the whole catalog to a language model?
- Navlu retrieves a small set of candidates first and only evaluates those, so input size does not grow with catalog size. Average response time is about 1.5 seconds.
- How does pricing work?
- There is no subscription and no mandatory monthly fee. You load usage credits and pay per use; usage is visible live in the dashboard. $100 for about 4,000 product searches. Search volume varies with conversation context and the number of products evaluated.
- How do I connect my products?
- Provide your catalog as JSON, CSV, XML or a read-only PostgreSQL view using the same field names. HTTPS feeds can update automatically on a schedule.
- Can I customize the assistant for my store?
- Yes. Colors, product cards, the mascot and which requests the assistant answers are configured per project. Installation is a single code snippet.
- Which languages does Navlu support?
- The product and assistant work in Turkish and English.