Multibrain
One workspace for 400+ AI models, billed by the credit.
- Years
- 2025
- Type
- Professional
- Discipline
- Web Dev, Full Stack, AI
- Stack
- Next.js, Typescript, Python, FastAPI, Supabase, DodoPayments
- Links
- Visit ↗
Multibrain is a unified AI access platform designed to simplify interaction with a wide range of language models through a single interface. Built as a wrapper over OpenRouter, the system provides structured access to more than 400 models while abstracting away cost management, concurrency, and persistence concerns.
The platform implements a credit-based usage system that dynamically calculates request costs based on model pricing, ensuring transparent and predictable billing for users. To support advanced workflows, the system enables concurrent execution of up to six models in parallel, while also allowing focused single-model interactions when required.
A persistent chat system maintains conversation history across sessions, with full Markdown rendering support for improved readability of model responses. Payments and wallet management are integrated directly into the platform using DodoPayments, enabling seamless credit top-ups and usage tracking.
The backend architecture was migrated from Next.js API routes to FastAPI, improving performance, scalability, and long-term maintainability. Supabase was used for database management, handling user data, conversations, and usage metadata.
What I did
- 01Built a unified platform providing access to 400+ AI models via OpenRouter
- 02Implemented a credit-based billing system with per-request cost calculation
- 03Designed concurrent execution pipelines supporting up to six parallel model queries
- 04Implemented persistent chat history with Markdown-rendered AI responses
- 05Integrated DodoPayments for secure payments and credit management
- 06Migrated backend services from Next.js to FastAPI for improved scalability and separation of concerns
(Inside)
01
Multi-Model Chat Interface
Unified chat interface enabling users to interact with 400+ AI models, with support for persistent chat history and markdown-rendered responses.
02
Concurrent Model Execution
Parallel execution pipeline allowing up to six AI models to be queried concurrently, with support for focused single-model interactions.
03
Credit-Based Usage & Billing
Dynamic credit deduction system calculating per-request costs based on model pricing and usage patterns.
04
Payments & Wallet Management
Integrated payment flow using DodoPayments for secure credit top-ups and usage-based billing.
05
Backend Architecture & Migration
Backend services migrated from Next.js API routes to FastAPI for improved scalability, performance, and separation of concerns.