LLM-Powered Chat Web App
Budget: ₹12,500 – ₹37,500 INR
This project entails building a generative-AI web application that runs entirely on OpenAI GPT and delivers a smooth, real-time, text-based chat experience. The frontend must be responsive so users on desktop or mobile feel no friction, while the backend handles dynamic prompts, keeps short-term conversational context, and lets me fine-tune how the model responds.
Core functionality
• Secure user authentication (log in / sign-up)
• Drag-and-drop or button-based file sharing inside the chat window
• Conversational memory so the AI remembers context during a session and resets cleanly when required
Behind the scenes the app should:
• Call the OpenAI GPT API efficiently, batching or streaming tokens where it makes sense to keep latency low
• Store transient chat history in a lightweight database or in-memory store for quick retrieval
• Offer simple control over system, user, and assistant prompt templates so I can adjust tone or domain-specific guidance without redeploying code
Deliverables
1. Source code for both frontend and backend, clearly organised and documented
2. Deployment instructions (Docker or similar) so I can launch on a standard cloud host
3. A short README explaining how to obtain and add my own OpenAI API key, plus configuration for file-size limits and session time-outs
4. Demo URL or video walkthrough showing the chat, authentication flow, and a file upload in action
Acceptance criteria: the chat must authenticate a user, let that user drop in a file, reference or summarise the file content on request, and respond in under five seconds for an average prompt.
Core functionality
• Secure user authentication (log in / sign-up)
• Drag-and-drop or button-based file sharing inside the chat window
• Conversational memory so the AI remembers context during a session and resets cleanly when required
Behind the scenes the app should:
• Call the OpenAI GPT API efficiently, batching or streaming tokens where it makes sense to keep latency low
• Store transient chat history in a lightweight database or in-memory store for quick retrieval
• Offer simple control over system, user, and assistant prompt templates so I can adjust tone or domain-specific guidance without redeploying code
Deliverables
1. Source code for both frontend and backend, clearly organised and documented
2. Deployment instructions (Docker or similar) so I can launch on a standard cloud host
3. A short README explaining how to obtain and add my own OpenAI API key, plus configuration for file-size limits and session time-outs
4. Demo URL or video walkthrough showing the chat, authentication flow, and a file upload in action
Acceptance criteria: the chat must authenticate a user, let that user drop in a file, reference or summarise the file content on request, and respond in under five seconds for an average prompt.