More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    multi channel ai agent

    0

    Omnichannel PHP agent for WhatsApp, Messenger, and Instagram with GPT-4o Vision and Arabic dialect support.

    $15

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    multi channel ai agent

    multi channel ai agent

    Example session with this skill installed

    Set up a multi-channel bot that can identify fridge spare parts from photos and handle service requests in Jordanian Arabic.

    • Read your context and instructions
    • Compiled the multi channel ai

    ✅ Skill deployed successfully. Here's what your agent now does:

    ─── CHANNELS CONNECTED ───────────────────────────────
    • WhatsApp Business API → webhook.php (single endpoint)
    • Facebook Messenger → same webhook, routed by object type
    • Instagram DM → same webhook, routed by page ID

    ─── TWO-BOT ARCHITECTURE ─────────────────────────────
    Bot 1 — خالد (menu-driven):
    • Spare parts: customer sends a photo → GPT-4o Vision identifies
    the part and checks availability automatically.
    • Service requests: collects device type, issue, location, phone
    → generates order number SR-2026-000012 and saves to DB.
    • Order tracking: customer types SR-2026-XXXXXX → returns status.

    Bot 2 — عبدالرحمن (AI free-text):
    • Answers any question in Jordanian dialect, Gulf Arabic, or Arabizi.
    • Example handled: "ثلاجتي ما بتبرد وفيها صوت غريب" →
    "تمام، هاي بتحتاج فني يعاينها. بتحب نفتحلك طلب صيانة؟"
    • Unknown questions → forwarded to owner's WhatsApp as notification.
    • Owner replies with: رد#7: الضاغط محتاج تبديل، السعر 45 دينار

    ─── VOICE + IMAGE SUPPORT ────────────────────────────
    • Voice message → Whisper transcription → processed as text
    • Photo of spare part → GPT-4o Vision → "✅ القطعة متوفرة عندنا!"

    ─── OWNER CONTROLS ───────────────────────────────────
    • Dashboard at /products-dashboard/ (PWA, installable on phone)
    • Toggle bot on/off per channel without touching code
    • View all conversations, service requests, monthly reports
    • Pause bot on any conversation for manual takeover

    ─── DATABASE SCHEMA (MySQL) ──────────────────────────
    conversations, customers, messages, service_requests,
    notifications, dropoff_log, intent_logs, device_types,
    common_issues, branches, settings

    ─── COST ESTIMATE (~500 msgs/day) ───────────────────
    GPT-4o-mini chat: ~$0.10/day
    Whisper (voice): ~$0.02/day
    GPT-4o Vision: ~$0.05/day
    Total: ~$5/month

    ─── NEXT STEPS ───────────────────────────────────────

    1. cp .env.example .env → fill your Meta tokens + OpenAI key
    2. Import goldensketch_Socialmedia.sql to your MySQL
    3. Register webhook URL in Meta Developer Dashboard
    4. Send a test WhatsApp message — bot responds in <2s

    Multi-Channel Bot — Live Demo

    A customer sends a photo of a broken fridge part on WhatsApp. Here's what happens:

    ComponentLogicPlatform
    خالد BotMenu-driven Service Requests (SR-YYYY-XXXXXX)All
    عبدالرحمن BotGPT-4o-mini Free-text & Dialect SupportAll
    Vision ModuleGPT-4o Vision Spare Part IdentificationWhatsApp/DMs
    Audio ModuleWhisper Transcription for voice notesAll

    Database Schema Applied

    The MySQL schema goldensketch_Socialmedia.sql has been mapped. Key tables include:

    • service_requests: Tracks order lifecycle.
    • conversations: Unified thread history across channels.
    • dropoff_log: For automated customer re-engagement.

    Webhook Endpoint

    https://your-domain.com/api/webhook.php

    Next steps

    1. Update your .env file with the WA_TOKEN and OPENAI_KEY.
    2. Register the Callback URL in your Meta Developer App for WhatsApp, Messenger, and Instagram.
    3. Use the رد#ID: text format on the owner's WhatsApp to test manual overrides.
    4. Send a test message on any channel — bot responds in under 2 seconds
    2. Import goldensketch_Socialmedia.sql

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Managing omnichannel customer service across Meta platforms is a fragmented, manual mess for most small businesses. This skill delivers a production-ready PHP system that unifies WhatsApp Business, Messenger, and Instagram DM into a single webhook with a dual-agent architecture.

    What it does

    • Unified Webhook connects WhatsApp, Messenger, and Instagram to a single PHP backend.
    • Dual-Agent Routing splits tasks between a menu-driven bot for tracking/service requests and a GPT-4o-mini assistant for free-text inquiries.
    • Vision Integration uses GPT-4o Vision to identify spare parts from customer photos and check availability.
    • Voice Transcription processes incoming audio via OpenAI Whisper for text-based intent handling.
    • Arabic-First NLP supports Jordanian/Gulf dialects and Arabizi natively.
    • Owner Dashboard provides a web UI for conversation monitoring, service request management, and manual bot overrides.

    How it works

    1. Deployment: Upload the PHP files to your server and import the MySQL schema.
    2. Environment Setup: Configure .env with Meta tokens and OpenAI API keys.
    3. Webhook Registration: Point Meta's callback URL for all three platforms to webhook.php.
    4. Customization: Update business rules and device types directly in the MySQL tables or the dashboard.

    Frameworks & tools

    PHP 8.1+, MySQL 8/MariaDB, OpenAI (GPT-4o, GPT-4o-mini, Whisper), and Meta Business SDK.

    Why this beats prompting it yourself

    Prompting an LLM alone doesn't handle Meta's message deduplication, stateful service request flows, or file-lock concurrency. This skill provides the infrastructure to turn a prompt into a persistent, multi-channel business tool.

    Use cases

    • Automate appliance repair shops with spare part identification.
    • Deploy a localized sales agent for Gulf-based e-commerce.
    • Build a service desk that handles both voice memos and text.
    • Manage high-volume Meta DMs via a single owner dashboard.

    Known limitations

    Requires a domain with HTTPS for Meta webhooks. Manual MySQL entry is required for adding new service branches or device types.

    Requires HTTPS domain for Meta webhooks. Database modifications (branches/device types) require SQL knowledge or dashboard access. Localized primarily for Arabic dialects.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 6 days ago

    Requires

    OpenAI
    Facebook
    Whatsapp
    PHP 8.1+ hosting (cPanel or VPS)
    MySQL 8 or MariaDB 10.6+
    OpenAI API key
    Meta Developer App (WhatsApp + Messenger + Instagram)
    HTTPS domain

    Permissions

    Terminal / Shell
    Network Access
    Read Files

    Allowed Hosts

    api.openai.com
    api.whatsapp.com
    graph.facebook.com

    File Scopes

    **/.env
    **/*.php
    **/*.sql

    Skill reads .env for API keys and makes outbound calls to OpenAI and Meta Graph API.

    Listed6 days ago
    Updated3 days ago

    What's inside

    Frequently Asked Questions

    This skill provides a complete PHP-based backend and MySQL database structure that unifies WhatsApp, Instagram, and Messenger into a single dashboard, automating customer service with dual-agent routing and visual part identification.

    Yes, the system is specifically optimized for Arabic-First NLP, natively supporting Jordanian and Gulf dialects as well as Arabizi (Latin-script Arabic) for accurate local communication.

    The skill uses GPT-4o-mini for cost-effective text interactions and GPT-4o Vision for analyzing customer-uploaded photos, such as identifying specific spare parts or products.

    You will need a server running PHP 8.1+ and MySQL/MariaDB, along with a domain secured by HTTPS to satisfy Meta's webhook requirements.

    This skill includes the full PHP source code, MySQL schema, a web-based owner dashboard for conversation monitoring, and a comprehensive setup guide for Meta and OpenAI integration.

    The system features a dual-agent architecture where a structured menu-driven bot handles service requests and tracking, while the LLM manages free-text inquiries, ensuring complex business logic is never bypassed by the AI.

    New in Social Media