Avito Card Parser Core

Python CLI core for collecting public Avito card data from search pages, seller profiles, or a prepared list of card URLs.

The main optimization is simple: collect card data from the HTML response and embedded page data instead of opening every card in a full browser. This avoids loading images, fonts, analytics, and client scripts for every item.

What It Does

• accepts a search/profile URL or a TXT file with card URLs;
• collects card links from result/profile pages;
• opens every card with lightweight HTTP requests, without Selenium/Playwright;
• extracts public fields and maps them to the report columns from the sample Google Sheet;
• saves progress and avoids duplicates;
• supports HTTP mobile proxy config and rotation URL;
• writes CSV immediately and can append to Google Sheets when credentials are provided.

Why This Is Faster

A browser-based scraper can load megabytes of extra static assets per card. This parser requests only the card HTML and parses the fields locally.

In a short local benchmark on 8 public cards, all 8 returned ok; the measured technical speed with a 1-2 second delay was about 850 cards/hour. For production-like use on one IP, the safer target is much lower: 90-100 cards/hour.

Config

Copy config.example.json and edit proxy / Google Sheets settings:

copy config.example.json config.json

Important fields:

• source_url: search results or seller profile URL.
• limit: number of cards to collect, up to 5000.
• cards_per_ip: rotate after this many parsed cards, normally 90-100.
• rotate_url: mobile proxy IP rotation URL.
• http_proxy / https_proxy: HTTP mobile proxy.
• google_sheets: service account settings.

Run

From search/profile URL:

python run_parser.py --config config.json --source "https://www.avito.ru/krasnodar/..." --limit 100 --resume

From a list of card URLs:

python run_parser.py -i examples/urls.txt -o out/avito_report.csv --limit 20 --min-delay 36 --max-delay 42 --resume

Short benchmark:

python run_parser.py -i examples/urls.txt -o out/test_report.csv --limit 8 --min-delay 1 --max-delay 2

Report Format

The CSV report uses the same column layout as the provided Google Sheets example:

Позиция, Категория, Комплектующие -> Ножки, Заголовок, Пр.Всего, Пр.Сегод., Цена, Описание, № объяв., Характеристики, Продавец, Широта, Долгота, Город, Ссылка, Фото шт., фото1, фото2, фото3, and other characteristic columns.

Notes

This project does not bypass captcha, access restrictions, authentication, or paywalls. If Avito returns a restriction page, the parser records the status, rotates IP when configured, and continues from the same queue.

Run checks:

python -m compileall .
python -m unittest discover -s tests

Portfolio Summary

Built a lightweight Avito card parser that extracts product data from HTML without full browser rendering. Added adaptive delays, progress saving, CSV/JSON export, and speed measurement for stable long-running collection.