AI Marketplace Monitor

An AI-based tool for monitoring facebook marketplace

• GitHub repo: https://github.com/BoPeng/ai-marketplace-monitor.git
• Documentation: https://ai-marketplace-monitor.readthedocs.io
• Free software: MIT

This program

1. Starts a browser (can be in headless mode)
2. Search one or more products
3. Notify one or more users of new products with phone notification

Features

• Search for one or more products using specified keywords.
• Limit search by minimum and maximum price, and location.
• Exclude irrelevant results.
• Exclude explicitly listed spammers.
• Exclude by description.
• Exclude previously searched items and only notify about new items.
• Send notifications via PushBullet.
• Search repeatedly with specified intervals in between.
• Add/remove items dynamically by changing the configuration file.

TODO:

• Use embedding-based algorithms to identify likely matches.
• Use AI to identify spammers.
• Support other notification methods.
• Support other marketplaces.

NOTE: This is a tool for programmers, and you are expected to know some Python and command-line operations to make it work. There is no GUI.

Quickstart

Install ai-marketplace-monitor

Install the program by

pip install ai-marketplace-monitor

Install a browser for Playwright using the command:

playwright install

Set up PushBullet

• Sign up for PushBullet
• Install the app on your phone
• Go to the PushBullet website and obtain a token

Write a configuration file

A minimal example is provided as minimal_config.toml. Basically you will need to let the program know which city you are searching in, what item you are searching for, and how you want to get notified.

[marketplace.facebook]
username = 'username'
password = 'password'
search_city = 'houston'

[item.name]
keywords = 'search word one'

[user.user1]
pushbullet_token = 'xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx'

A more complete example is provided at example_config.toml, which allows for more complex search and notification patterns. Briefly:

• marketplace.facebook allows

  • username: (required)
  • password: (required)
  • login_wait_time: (optional), time to wait before searching in seconds, to give you enough time to enter CAPTCHA, default to 60.
  • search_interval: (optional) minimal interval in minutes between searches
  • max_search_interval: (optional) maximum interval in minutes between searches
  • search_city: (optional if defined for item) search city, which can be obtained from the URL of your search query
  • acceptable_locations: (optional) only allow searched items from these locations
  • exclude_sellers: (optional) exclude certain sellers by their names (not username)
  • min_price: (optional) minimum price.
  • max_price: (optional) maximum price.
  • notify: (optional) users who should be notified for all items

• user.username where username is the name listed in notify

  • pushbullet_token: (rquired) token for user

• item.item_name where item_name is the name of the item

  • keywords: (required) one of more keywords for searching the item
  • marketplace: (optional), can only be facebook if specified.
  • exclude_keywords: (optional), exclude item if the title contain any of the specified words
  • exclude_sellers: (optional, not implemented yet) exclude certain sellers
  • min_price: (optional) minimum price.
  • max_price: (optional) maximum price.
  • exclude_by_description: (optional) exclude items with descriptions containing any of the specified words.
  • notify: (optional) users who should be notified for this item

Run the program

Start monitoring with the command

ai-marketplace-monitor

or

ai-marketplace-monitor --config /path/to/config.toml

NOTE

1. You need to keep the terminal running to allow the program to run indefinitely.
2. You will see a browser firing up. You may need to manually enter any prompt (e.g. CAPTCHA) that facebook asks for authentication in addition to the username and password that the program enters for you. You may want to click "OK" to save the password, etc.

Advanced features

• A file ~/.ai-marketplace-monitor/config.yml, if it exists, will be read and merged with the specified configuration file. This allows you to save sensitive information like Facebook username, password, and PushBullet token in a separate file.
• Multiple configuration files can be specified to --config, which allows you to spread items into different files.

Credits

• Some of the code was copied from facebook-marketplace-scraper.
• This package was created with Cookiecutter and the cookiecutter-modern-pypackage project template.