Hi everyone, I'm currently exploring how tools like Boardy.ai technically work under the hood. From the outside, Boardy.ai seems to integrate with platforms like LinkedIn and automate outreach or data collection through AI-driven interactions. I'm really curious about the architecture and stack choices behind such a system, and I'm hoping the community can help me break it down. Here are a few questions I’m trying to understand: 1. How might Boardy.ai connect to LinkedIn, given LinkedIn’s API restrictions? Are they using browser automation, cookies/session injection, or third-party proxies? 2. What kind of backend structure (Node.js, Python, serverless?) is suitable for managing user sessions, message parsing, and AI integration? 3. How is AI (like GPT) embedded into the user flow — real-time vs queued responses? 4. Do you think services like Vapi.ai or Puppeteer are part of the pipeline for voice or browser-based automation? 5. How would one handle scalability and data privacy concerns, especially when dealing with personal LinkedIn messages? If anyone has experience building LinkedIn-integrated tools or automation platforms like this, I’d love to hear your thoughts or see architecture examples. Thanks in advance!

I made LinkedIn scrap workflow in n8n, but I want to send messages to the profiles I pull in order, I want to do this just to try it, my goal is not to spam messages, how do you think I can realize this?