IGFlow — Instagram Automation Platform

About Project

IGFlow is an automation platform I built for businesses and agencies that run Instagram accounts. I designed and developed all of it, from the frontend to the server infrastructure. Live at https://igflow.app What it does • Follow, like, comment, story-view and data-collection tasks, throttled by hourly, daily and weekly limits that protect the account • Visual flow editor: drag-and-drop automations that react to DMs, mentions or new followers • Real-time inbox: read and answer DMs from the panel; automation pauses automatically when a human takes over • AI replies: agents that run on the user's own AI connection and MCP tools • Scheduled posts, Reels and stories • Plans and subscriptions, admin panel, support tickets, 5 languages Architecture The system is a set of independent services: a Next.js panel, a Spring Boot API, a worker that takes jobs over Kafka, a real-time gateway that keeps persistent connections to Instagram, and a small Go egress service that isolates outbound AI calls. Long-running jobs run in per-account ordered queues while different accounts run in parallel. Security and infrastructure • Two-server setup; the data layer (MongoDB, Redis, Kafka) is reachable only over a private WireGuard network • Secrets in HashiCorp Vault, sensitive fields encrypted with AES-256-GCM • JWT + HttpOnly sessions, TOTP two-factor auth, audit log, rate limiting, idempotency keys • Gitea Actions: tests and automatic deploy on every push Note: IGFlow uses session-based access rather than the official Meta API; this risk is stated clearly to users, and the system is designed to protect accounts with limits.