Build & Framework Details

AI Experiment BD is built as a modern, scalable, web-based agentic creative platform, combining a responsive React interface with generative AI services and a WebMCP tool layer that allows AI agents to interact directly with application capabilities.

Frontend

• Next.js — application framework and production web runtime
• React — component-based UI architecture
• TypeScript — type-safe application and tool development
• Tailwind CSS — responsive design system and utility-based styling

AI & Creative Layer

• Generative AI APIs for multimodal creative workflows
• Prompt enhancement and structured prompt generation
• AI image generation and editing
• Bengali/English text generation
• Story and storyboard generation
• AI voice/narration workflows

WebMCP / Agent Layer

The core hackathon innovation is the WebMCP integration.

AI Experiment BD exposes application capabilities as structured WebMCP tools using the browser's Model Context interface. These tools allow compatible AI agents to discover and operate the creative workspace without relying exclusively on fragile UI navigation.

Example capabilities include:

• create_project
• enhance_prompt
• generate_image
• edit_image
• generate_story
• generate_storyboard
• generate_voice
• list_project_assets
• save_project_asset
• export_project

Application Architecture

Human → Creative UI → Project State → WebMCP Tool Layer → AI Agent → Creative Services

The application maintains structured project context so an agent can understand the current task, inspect available assets, perform multi-step operations, and iterate based on human feedback.

Backend & Data

• Server-side API routes for AI operations
• Secure environment-variable configuration for API credentials
• Structured project and asset state
• Modular service architecture for future AI providers
• Validation and error handling around agent tool execution

Deployment

The application is designed for production web deployment with:

• HTTPS
• Environment-based configuration
• Responsive desktop/mobile experience
• Public live demo
• Public open-source repository
• Reproducible development and deployment workflow

Open-Source Architecture

The project is intentionally modular so developers can extend the WebMCP tool registry with additional creative capabilities, connect new AI providers, and build their own agent-driven workflows.

Core principle: AI Experiment BD is not simply an AI website with a chatbot. It is a tool-enabled creative environment where WebMCP gives AI agents a structured interface to the application's real capabilities.

Built With

• agentic
• agents
• ai
• bengali
• css
• engineering
• generation
• generative
• image
• mcp
• multimodal
• next.js
• prompt
• react
• rest
• storytelling
• tailwind
• typescript
• voice
• webmcp
• workflows