We build e-commerce platforms for businesses

That's the first thing we're discussing. We've moved it from WordPress, where there were about 53,500 addresses in the Yandex index. The mechanism is this: we compare old and new URLs by article, we give live goods 301 to the exact address, we give the removed ones 410 Gone, we give categories without a direct pair 301 to the relevant section. We don't do a mass redirect to the main page: Yandex considers it soft-404 and lowers the site. For this to work, we need to download the address of the article from the database of your current website.

Imports to the site work automatically: catalog, categories, brands, residues and personal prices come from 1C without manual labor. When we're sending orders back into account, we're setting them to fit your contour -- the 1C configuration, the processor, and the exchange format are different for everyone, so this is a separate stage of work, not a setup glitch. At the same time, the manager instantly learns about the order in a convenient channel - a sound push in CRM, mail, Telegram, MAX or VK - and changes the status in a couple of buttons, so no request is lost, even while the exchange is set up.

At the core of the platform are a web showcase and Telegram - Mini App for the buyer and a bot for the manager. Mini Apps for VK and MAX, as well as bots for them, we've implemented in our own product Razdaczka, where they work on the edge, and we connect to the platform with a separate module at the implementation stage. API-first architecture: any new channel is going to the same core as the site, so the directory, prices, and bonuses are the same.

No, I'm the only one. Accounts are glued to a phone number: the phone is assigned to a single format and serves as an identification key, so the profiles from Telegram, VK, MAX, and the site are aggregated into one user. The buyer has one cashback balance, one order history, and one delivery address in each channel, and you have one customer card instead of three. If the automatic matching doesn't work, the operator manually connects the profiles from the admin.

And then manual cataloging, weak B2B and lack of retention cost money every day - missed orders and time for managers. If you're unfamiliar with these pains, there's really nothing to change. We'll be honest about it at the audition.

We're not writing the platform from scratch - we're rolling out the finished product to you individually, and with each implementation, it's cheaper. Plus, it's modular: you can start with one block -- account integration, loyalty programs, or Telegram sales -- and build the rest gradually, instead of paying for everything at once.

The platform's logic was originally designed for a complex 1C nomenclature: a hierarchy of categories, brands and product groups, thousands of SKUs and B2B pricing with personal prices for each organization. The images are uploaded by the archive packet through the article and automatically converted to WebP.

The paywall we connect to the client at the implementation stage is not a pre-designed function. For B2B, there's already an account payment at the core with the organization's details and the choice of bank account.

Now the platform is giving YML feed to Yandex.Market. Fully functional two-way connectors to marketplaces are a separate task for a particular platform; API-first architecture allows this, but we won't promise integration with marketplaces out of the box.

That's right. The platform is written on an open stack - Laravel, Nuxt, PostgreSQL - rather than a closed box. The code and data are yours, the documentation is transmitted, any team that knows this stack can support the project.

The platform is deployed on two combat instances: Vordeke and Ma La Skin. It's a single code base for two different clients -- that is, it's not a presentation that's verified, it's a proofreading. At the same time, each instance is updated according to its schedule: the second client does not automatically receive everything that appeared in the first.