Corner Shop Chat
Status: Live.
Your muse
Rebuild ofBuilding a multimodal WhatsApp chatbot with Llama 4 (opens in a new tab)Partial
Make a customer chat for a small shop that understands photos and voice notes. This rebuilds Meta's video "Building a multimodal WhatsApp chatbot with Llama 4" (https://dev.meta.ai/llama/resources/videos/build-multi-modal-whatsapp-chatbot-with-llama-4) on current Muse models. It demonstrates: A business chatbot on WhatsApp that handles text and images. Plan: A web chat for a small business that understands photos and voice notes and can reply with images. Connecting it to WhatsApp needs a WhatsApp Business account, so the rebuild shows the web version and documents the webhook.
Checks passed
The plan: Corner Shop Chat
A web chat for small shops that answers text, photos and voice notes.
Can build part of it
- Owner enters shop name, hours and product list
- Customer starts chat and sends text, a photo or a voice note
- App transcribes voice and understands text plus photo
Won't do: Rebuild fit: partial
See the full plan
Who it's for
Small shop owners and their customers
How it works
- Owner enters shop name, hours and product list.
- Customer starts chat and sends text, a photo or a voice note.
- App transcribes voice and understands text plus photo.
- App replies in text and adds a generated image when asked.
- Customer reads answer and owner sees webhook setup guide.
Screens
- Shop setup. Owner adds shop details, hours and products to guide answers.
- Customer chat. Customer sends text, photos and voice notes and gets answers with images.
- Connect guide. Shows how a developer would move the web chat to WhatsApp with a webhook.
What goes in
- Text: Shop profile with name, hours and products, plus customer text message.
- Image (optional): Optional customer photo of a product or issue, JPG or PNG under 4.5 MB.
- Audio (optional): Optional short voice note. Browser converts to mono 16-bit WAV before transcription.
What comes out
- Text: Friendly shop answer that uses the photo and voice transcript when provided.
- Image: Optional generated image showing a product example or simple visual help.
Models
Model Used for Why Voice Transcribe Turn customer voice notes into text Needed to understand spoken questions. Converts browser recorded WAV into a plain transcript. Muse Spark 1.3 Understand text, photos and transcripts and write shop replies Default reasoning and vision model. Reads photos and text together and returns helpful answers. Muse Image Create reply images for products and examples Needed to reply with images. Generates a simple shop image when the customer asks to see an item. Limits
- Rebuild fit: partial. Original video shows a WhatsApp chatbot with Llama 4. This rebuild is web only because products cannot reach WhatsApp and it only documents the webhook.
- No live WhatsApp connection. It shows setup steps and webhook docs but cannot send messages or use a WhatsApp Business account.
- Voice notes must be short recorded files converted to mono 16-bit WAV. No live calls and no speaker labels.
- Photo answers and generated images are estimates. Hidden details can be missed and reply images are illustrations, not exact stock.
From your inspo
Rebuild of the Meta developer video Building a multimodal WhatsApp chatbot with Llama 4. It keeps the small shop chat that understands text, photos and voice notes and can reply with images, as a web version.
Takes
- Small business chatbot job that answers customer questions
- Understanding text plus product photos in one chat
- Simple question and answer flow for a shop
Leaves out
- Video footage
- Video script and spoken words
- Meta and Llama names and logos
- WhatsApp brand and interface
Cost per run: about $0.01 (est.). Voice transcription about $0.001 plus Spark chat with photo about $0.004 plus one reply image $0.01 equals about $0.015 per run.
Recorded demo Done. Built Corner Shop Chat and
npm run verifyexits 0 (all steps passed, 5 tests passed).