Multilingual Viber Ads Auto-Poster
Budget: €30 – €250 EUR
I run www.MarcheTamil.fr and need a small application that sits between my Viber community and the site. The tool must listen to every new message that appears in the group (or is sent to an associated Viber bot; I want both entry points covered), decide whether the text is truly a classified ad, and—only if it includes a phone number—extract the key data, file it under the correct category, and push it to my site where I can approve it before it goes live.
Language handling is critical: ads may arrive in Tamil, French or English, so the detection and field-extraction logic has to work seamlessly across all three. For every genuine ad the app should pull out the title, full description, phone number, location, price and any attached images, check for duplicates, then post through my site’s API. Anything that is missing a phone number or is obviously not an ad must be ignored.
I would like:
• A lightweight dashboard showing pending, approved and rejected ads
• Manual “approve / reject / edit” controls prior to publication
• Clean, secure, well-commented code that I can deploy on my own server
Please propose the tech stack you feel is best suited (Python, Node, or similar—open to suggestions) along with an outline of how you will tackle multilingual NLP and Viber integration.
Language handling is critical: ads may arrive in Tamil, French or English, so the detection and field-extraction logic has to work seamlessly across all three. For every genuine ad the app should pull out the title, full description, phone number, location, price and any attached images, check for duplicates, then post through my site’s API. Anything that is missing a phone number or is obviously not an ad must be ignored.
I would like:
• A lightweight dashboard showing pending, approved and rejected ads
• Manual “approve / reject / edit” controls prior to publication
• Clean, secure, well-commented code that I can deploy on my own server
Please propose the tech stack you feel is best suited (Python, Node, or similar—open to suggestions) along with an outline of how you will tackle multilingual NLP and Viber integration.