Telegram Bot: Photos → Direct Links → Google Sheet (AI Vision) — 1-Day
Budget: €30 – €250 EUR
Serious freelancers only — 1-day delivery (max +2 days if agreed).
Goal: Telegram bot to receive product photos, save them on my domain as direct public links, and write one row to a Google Sheet. AI Vision suggests Title, Description, Type/Category, Color (Brand if confident). I can set Brand manually.
User flow: /new → bot auto-IDs (YYYY-####, or I send #ID) → send photo(s) (first = main) → send discounted price then full price (supports 750/1000, currency symbols, -25%) → /save.
Data columns: Product ID | Main Photo | Additional Photos | Title | Description | Type (L1) | Category (L2) | Color | Gender (default M; /gender F|K) | Brand (my input has priority) | Supplier/Warehouse (default empty; /supplier …) | Full Price | Discounted Price | Needs Review (TRUE/FALSE).
Type/Category/Color must be from controlled lists; if unsure → leave empty + Needs Review = TRUE.
You create/register & hand over: Google service account (Sheets API), HTTPS config (Nginx/Apache), /uploads structure, deployed bot (systemd/Docker), repo with full source + README, .env.example, keys/credentials, short admin guide.
Final Acceptance Test (my check): I upload ~20 items (multi-photo + different price formats). Pass if: all rows created with correct IDs; direct links open; prices parsed correctly (discounted ≤ full); vocab values only; gender/brand/supplier commands work; /status, /edit price, /save, /cancel OK; survives service restart.
You self-test first. I run one final test; up to 3 final attempts to fix. If not passed or deadline missed → project cancelled, no payment.
What I provide: domain/DNS or hosting access, Telegram bot token, Google Sheet (or permission to create), AI API key, my Telegram IDs, vocab lists.
Stack: Python (aiogram) or Node.js, Telegram API, Google Sheets API, Vision/OCR, Docker, Nginx/HTTPS.
Telegram API; Python; Google Sheets API; Google Cloud Platform; OpenAI API; Artificial Intelligence; Computer Vision (OCR); Docker; Nginx; Linux Administration
Goal: Telegram bot to receive product photos, save them on my domain as direct public links, and write one row to a Google Sheet. AI Vision suggests Title, Description, Type/Category, Color (Brand if confident). I can set Brand manually.
User flow: /new → bot auto-IDs (YYYY-####, or I send #ID) → send photo(s) (first = main) → send discounted price then full price (supports 750/1000, currency symbols, -25%) → /save.
Data columns: Product ID | Main Photo | Additional Photos | Title | Description | Type (L1) | Category (L2) | Color | Gender (default M; /gender F|K) | Brand (my input has priority) | Supplier/Warehouse (default empty; /supplier …) | Full Price | Discounted Price | Needs Review (TRUE/FALSE).
Type/Category/Color must be from controlled lists; if unsure → leave empty + Needs Review = TRUE.
You create/register & hand over: Google service account (Sheets API), HTTPS config (Nginx/Apache), /uploads structure, deployed bot (systemd/Docker), repo with full source + README, .env.example, keys/credentials, short admin guide.
Final Acceptance Test (my check): I upload ~20 items (multi-photo + different price formats). Pass if: all rows created with correct IDs; direct links open; prices parsed correctly (discounted ≤ full); vocab values only; gender/brand/supplier commands work; /status, /edit price, /save, /cancel OK; survives service restart.
You self-test first. I run one final test; up to 3 final attempts to fix. If not passed or deadline missed → project cancelled, no payment.
What I provide: domain/DNS or hosting access, Telegram bot token, Google Sheet (or permission to create), AI API key, my Telegram IDs, vocab lists.
Stack: Python (aiogram) or Node.js, Telegram API, Google Sheets API, Vision/OCR, Docker, Nginx/HTTPS.
Telegram API; Python; Google Sheets API; Google Cloud Platform; OpenAI API; Artificial Intelligence; Computer Vision (OCR); Docker; Nginx; Linux Administration