REAL PROJECT · IN PRODUCTION

Open-data collection service

Parser

A B2B platform for planning data collection, gathering public contacts, deduplicating results, monitoring runs and exporting datasets.

Open live project ↗parser.finansteh.ru
parser.finansteh.ru
Parser home page

PRODUCT OVERVIEW

What the service is and why it exists

Parser turns manual research into a managed workflow. A user can start with a free-form request or template, choose cities and sources, follow processing stages and receive structured records containing contacts, source links and working statuses.

TASK

Reduce repetitive manual business research by replacing site-by-site browsing, contact copying, duplicated searches and scattered result storage with one controlled workflow.

SOLUTION

We built a web service where users describe the target, choose sources and geography, review a parsing plan, run the collection and work with results in a persistent database. Run history makes successful searches repeatable.

WHAT WAS IMPLEMENTED

The product in detail

Core product capabilities that are already built into the working service.

01

Guided task creation

A natural-language request becomes a plan covering scope, search queries, extraction, deduplication and final storage.

02

Geography and batch runs

Searches can target one city, a region or a series of geographic segments for repeatable B2B prospecting.

03

Multiple source modes

Open-web, URL, sitemap and API workflows are supported, with an architecture designed for additional sources.

04

Contact database

Company name, city, phone, website, email, social links, address and source URL can be stored in one working dataset.

05

Deduplication and cleanup

Duplicate records from different searches and sources are merged to keep the final database usable.

06

History and templates

Runs are preserved and repeatable scenarios can be saved as templates instead of being configured again.

07

Progress monitoring

Users can see processing stage, cities covered, contacts collected and the current run state.

08

Export and API

Results can be exported for downstream work or consumed through API-oriented integration flows.

USER FLOW

From the first action to the result

  1. 01

    The user describes the target or chooses a prepared template.

  2. 02

    Geography, source modes and required fields are selected.

  3. 03

    The service builds a plan and splits the work into search stages.

  4. 04

    Collectors traverse permitted public sources and extract available data.

  5. 05

    Records are normalised and deduplicated.

  6. 06

    The final dataset is saved in the account for filtering, statuses, export and continued work.

INTEGRATIONS & AUTOMATION

Connected to the surrounding business process

  • Search and public web sources, URL/sitemap workflows and pluggable APIs.
  • User authentication and a persistent account with run history.
  • YooKassa for paid plans and subscription-oriented flows.
  • Internal notifications, customer administration and access-state control.

ENGINEERING

Built for production, not only for a demo

  • Task queue and staged execution for long-running collection jobs.
  • Field normalisation and deduplication across multiple sources.
  • Export flows and an API-ready data model for external systems.
  • Production deployment on a private VPS with HTTPS and automated releases.

RESULT

A working product with room to grow

The working SaaS product turns repeatable B2B research into a structured dataset that can be used directly for sales operations and analysis.

Web appParsing workflowsSearch sourcesDeduplicationExports & APIYooKassaUser cabinetVPS / Nginx

Need a product with a similar level of care?

Tell us about the goal. We can start with a landing page, a first product release or focused automation.

Discuss a project ↗

AVEXIUM chat

AI replies instantly · operator on request

AI
AVEXIUM AI
Hello! Tell me what you want to build and I’ll help you get oriented. Use “Transfer to operator” whenever you want a person.

After transfer, new messages go to the operator in Telegram. Replies will appear here.