BotLode

A comprehensive multi-platform ecosystem to build, embed, and monitor Gemini 2.0 Flash chatbots with automated lead scoring, real-time alerts, and an integrated CRM.

Comment

Inspiration

What it does## Inspiration

Most AI chatbot implementations are treated as isolated widgets rather than complete business solutions. Business owners don’t just need an AI that talks; they need an ecosystem that creates bots easily, embeds them seamlessly, tracks buyer behavior, and alerts them when a high-value lead is found. We built BotLode Suite to bridge the gap between advanced generative AI and a full-scale conversion and CRM pipeline.

What it does

BotLode Suite is an informal monorepo composed of 4 core pillars that operate end-to-end:

  • BotLode Factory (Windows Desktop): The administrative core built for Windows where users manage billing, configure global bot behavior, and deploy new assistants using predefined blueprints.
  • BotLode Player (Embedded Iframe): An ultra-lightweight web runner designed to be embedded in third-party websites via iframes. It features interactive Rive animated avatars that physically track the user's cursor dynamically.
  • BotLode History (The CRM): A dedicated web dashboard for bot owners to review chat logs, manage calendar availability, and book unique meeting slots.
  • Smart Pipelines: Every conversation runs through a real-time analytics layer that calculates an intent_score. If a hot lead is detected (Score >= 80), it instantly fires an automated email alert to the business owner.

How we built it

We designed a decoupled architecture connected by a single backend:

  • AI Core: Powered by Supabase Edge Functions (Deno) running Gemini 2.0 Flash with advanced multi-turn conversation memory and contextual summary windows.
  • Frontend Diversity: Built with Flutter (Desktop for the Factory Terminal, and HTML-renderer Web for the Player and CRM). The modern marketing landing page is being migrated to Next.js 16 + React 19 for SEO performance.
  • Backend & Integrations: Hosted entirely on Supabase for real-time databases and edge routing, integrated with Resend for lead notifications, Twilio Business API for message queues, and Mercado Pago for automated billing cycles.

Challenges we faced

Synchronizing cross-origin user behaviors was tough, especially implementing bi-directional postMessage communication between host websites and the Flutter Web player iframe to handle real-time Rive animation tracking based on cursor coordinates. We also spent significant effort fine-tuning prompt boundaries in Gemini 2.0 Flash to ensure stable, reliable intent-scoring without false positives.

What's next for BotLode Suite

Our immediate roadmap includes moving from single-tenant schemas to a secure multi-tenant architecture, hardening Row-Level Security (RLS) policies on critical CRM tables, consolidating our web presence into our Next.js frontend, and adding native WhatsApp/Telegram sync to the AI core.

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for BotLode

Built With

Share this project:

Updates