Budget: 6000 UAH Deadline: 3 days
Hello! 👋
I am ready to take on the refinement of your Telegram bot.
The task is clear: it is necessary to add the ability to recognize products from photos with subsequent generation of names and keywords — while maintaining the current logic of working with text.
Here’s how I see the implementation:
1. Image processing — the user attaches a photo, the bot receives it and sends it to an AI model (for example, Google Vision API, OpenAI Vision, or a similar one) to determine the category/name of the product.
2. Internet search — after recognition, the bot uses the existing logic to gather names and descriptions from marketplaces (Prom, Rozetka, Epicenter).
3. Result generation — I will form a structured list: product name, relevant keywords, search queries.
4. UX optimization — I will add image format verification, a processing indicator, and error handling.
Risks / nuances:
• The API for image recognition has limitations (rate limit or cost of requests) — I will find the optimal option with a free or local model.
• For quality product identification, adaptation to Ukrainian marketplaces will be needed (English-speaking models may confuse brands).
• I will check stability to ensure the bot does not "hang" when processing large photos.
Proposal:
⏱ Deadline: 3–5 days
💰 Cost: approximately $120–150, depending on the chosen API and desired level of accuracy.
If you have examples of photos for testing — I would appreciate it to check the accuracy of recognition right away.