BizThrive.ai
Workflows by BizThrive.ai
AI-powered invoice extractor via Telegram to Airtable
## 📄 Description This workflow automates the extraction of structured invoice data from PDF files sent via Telegram and stores it in Airtable. It leverages GPT-4o for intelligent parsing and includes conversational memory for a seamless user experience. Designed for businesses and freelancers who receive invoices digitally and want to streamline their record-keeping. --- ## ⚙️ How It Works 1. **Telegram Trigger** – Listens for incoming messages and PDF attachments. 2. **Switch Node** – Filters messages to ensure only PDFs are processed. 3. **Extract from File** – Parses the PDF content for text extraction. 4. **Edit Fields** – Prepares the extracted data for AI processing. 5. **AI Agent (GPT-4o)** – Orchestrates the workflow, prompts the user for missing info, and extracts structured data. 6. **Simple Memory** – Maintains conversational context across sessions. 7. **Create Invoice (Airtable Tool)** – Creates a new invoice record in Airtable. 8. **Create Line Item (Airtable Tool)** – Adds individual line items linked to the invoice. 9. **Telegram Response** – Sends confirmation back to the user. --- ## 🔐 Required Credentials To run this workflow successfully, you’ll need: - **Telegram Bot Token** (via @BotFather) - **OpenAI API Key** (with GPT-4o access) - **Airtable API Key** and access to: - Base: `Invoice Tracker Proper` - Tables: `Invoices` and `Line Items` --- ## 🧰 Airtable Structure **Invoices Table Fields**: - Invoice Number - Date - Supplier - Supplier Address - Tax ID - PO Number - Due Date - Receiver Name - Receiver Address - Delivery Date - Total Tax - Total Amount **Line Items Table Fields**: - Product Code - Description - Unit Price - Quantity - Unit Type - Sub Total - Invoice (linked) --- ## 🧠 Features - AI-powered invoice parsing - PDF text extraction - Airtable record creation with relational linking - Telegram-based user interaction - Conversational memory - Error handling and data validation
Deep research assistant with Perplexity AI and Telegram citations
**Turn your Telegram bot into a real-time research assistant** with this intelligent n8n workflow. Designed for founders, analysts, and knowledge workers, this automation uses Perplexity Sonar and Sonar Pro to deliver **concise, citation-rich answers** to complex queries — directly inside Telegram. --- **🔍 What It Does** - **✅ Smart Query Routing** Automatically selects the right tool based on query complexity — Sonar for fast lookups, Sonar Pro for multi-source synthesis. - **📚 Cited Research Summaries** Includes clickable URLs from Perplexity’s source data for transparency and auditability. - **🧠 Session Memory** Maintains chat context using Telegram chat ID for follow-up questions and threaded insight. - **🔐 Secure Access Filter** Restricts bot usage to authorized Telegram users. - **⚙️ Customizable Agent Behavior** Easily adjust tone, tool preferences, and citation style via system message. --- **🚀 Use Cases** - Market research & competitor analysis - Academic and scientific deep-dives - Legal and transcript summarization - Podcast, video, and trend monitoring - Personal AI assistant for founders and consultants --- **🛠 Setup Instructions** 1. Create a Telegram bot via [@BotFather](https://t.me/BotFather) and add your token. 2. Add your OpenAI and Perplexity API keys. 3. Update the filter node with your Telegram user ID. 4. Deploy and start chatting — responses appear in Telegram.