Case study
Botsify — Agentic AI Platform
Built the entire agentic AI layer for a no-code chatbot platform — multi-LLM orchestration, real-time voice AI, RAG pipelines, and MCP integrations.
Context
Botsify needed a complete agentic AI overhaul — moving from rule-based chatbots to LLM-powered agents with real-time voice, document RAG, and dynamic tool use via MCP.
Approach
Architected a unified LLMService abstraction supporting 6+ providers, built real-time voice pipelines with Twilio and Deepgram, implemented RAG with Cloudflare Vectorize, and designed an MCP layer for live agent tool connections.
Key features
- Multi-LLM provider engine with dynamic switching
- Real-time voice AI calls with interrupt handling
- MCP integration for agent tool connections
- RAG pipeline with Cloudflare Vectorize
- Vue 3 AI management dashboard (20+ views)
- AI Skills marketplace with Fly.io isolation
- Guardrails system for output compliance
- Whitelabel enterprise platform
<1s
Voice Latency
6+
LLM Providers
Responsibilities
- Architect multi-LLM provider engine with unified abstraction
- Build real-time voice AI with Twilio, Deepgram, and OpenAI
- Design MCP integration layer for live tool connections
- Implement RAG with Cloudflare Vectorize
- Build Vue 3 + TypeScript + Pinia AI dashboard
- Develop AI Skills marketplace on Fly.io
- Extend Laravel backend with agentic AI layer
- Build whitelabel enterprise platform
Outcomes
- Sub-1-second voice AI latency
- 6+ LLM providers supported dynamically
- Full MCP tool ecosystem for agents
- Document-grounded responses via RAG
- Enterprise-ready whitelabel deployments
Site preview
White Label AI Chatbot Platform | Botsify
Build and resell AI chatbots and AI agents under your own brand. Manage clients, customize your platform, and grow your agency with Botsify.
- Sell AI.To clinics.Get paid every month.
- The infrastructure behind your agency.
- Ready to build this under your brand?
