Prompt.me

Overview

Prompt.me is a multi-model AI SaaS platform that enables users to send one prompt to multiple large language models simultaneously and compare responses side by side.

Built with Next.js, TypeScript, Firebase, Clerk, Stripe, OpenAI, Anthropic Claude, and Google Gemini, the platform brings AI orchestration, independent response streaming, usage tracking, and subscription billing into one application.

Tech stack: Next.js, React, TypeScript, Tailwind CSS, Firebase, Firestore, Clerk, Stripe, OpenAI API, Anthropic Claude, and Google Gemini.

Problem Statement

1. Fragmented AI Model Access

Using multiple AI providers requires separate integrations and provider-specific request handling, making it difficult to compare responses through one consistent interface.

2. Slow or Unreliable Multi-Model Responses

Parallel model requests introduce challenges around independent streaming, provider timeouts, and API failures. One slow or unavailable provider should not block the remaining responses.

3. Complex AI Usage and Cost Tracking

Different models have different usage costs, requiring accurate token tracking, per-model pricing, balance validation, and credit deductions.

4. Subscription and Usage Management

The platform needed to coordinate user authentication, subscription billing, credit balances, conversations, comparison history, and usage records.

Solution Statement

1. Unified Multi-Provider AI Architecture

Built a shared AI provider abstraction for OpenAI, Anthropic Claude, and Google Gemini, normalizing model requests, response formats, streaming, errors, timeouts, token usage, and cost tracking.

2. Parallel Streaming and Provider Isolation

Implemented concurrent prompt execution with independent streaming pipelines, per-provider timeout handling, and failure isolation so one provider’s failure would not block successful responses from other models.

3. Credit-Based Usage Metering

Developed a usage system with per-model pricing, pre-request balance validation, token tracking, credit deductions, provider-level cost tracking, usage history, and model-specific cost visibility.

4. Integrated Authentication and Billing

Connected Clerk authentication, Stripe subscription billing, and Firestore data models to manage accounts, prompts, conversations, comparisons, model responses, usage records, and credit transactions.

Features

• Multi-model AI comparisons using a single prompt.
• Side-by-side responses from OpenAI, Claude, and Gemini.
• Concurrent AI requests with independent response streaming.
• Unified provider abstraction for model requests and response formats.
• Provider-specific timeout handling and failure isolation.
• Model selection for individual comparisons.
• Token usage tracking and provider cost visibility.
• Credit-based usage metering and pre-request balance validation.
• Stripe subscription and payment integration.
• Clerk authentication and account management.
• Firestore persistence for prompts, conversations, comparisons, and usage.
• Responsive desktop and mobile interface.
• API usage and provider error monitoring.

Benefits

• Simplified AI Comparison: Users can submit one prompt to multiple models and review responses in a single interface.
• Responsive Model Results: Independent streaming allows responses to appear without waiting for every provider to finish.
• Improved Fault Isolation: Provider-level timeouts and error handling help prevent one failed integration from blocking others.
• Greater Cost Visibility: Token tracking and model-specific cost information make AI usage easier to monitor.
• Controlled Credit Usage: Balance validation and credit deductions connect model usage with the platform’s metering system.
• Integrated Subscription Management: Stripe billing works alongside the internal credit system.
• Structured Conversation History: Firestore models support stored prompts, comparisons, responses, and usage records.
• Maintainable AI Integrations: A unified provider layer reduces the need to spread provider-specific logic throughout the application.

My Role

As the Full Stack Engineer, I owned frontend development, backend implementation, AI integrations, billing workflows, data architecture, and production delivery.

My responsibilities included building the Next.js application, architecting multi-provider orchestration, implementing parallel streaming and provider isolation, designing Firestore data models, integrating Clerk and Stripe, developing credit-based usage metering, and setting up CI/CD and production monitoring.

Core Expertise Demonstrated

Full-stack development, AI SaaS development, Next.js, React, TypeScript, Tailwind CSS, OpenAI API, Anthropic Claude, Google Gemini, LLM orchestration, parallel processing, streaming responses, Firebase, Firestore, Firebase Admin SDK, Clerk authentication, Stripe integration, subscription billing, credit-based billing, usage metering, token tracking, cost optimization, serverless architecture, CI/CD, and API monitoring.