What is ModelStream?

ModelStream aggregates multiple upstream AI providers (OpenAI, Claude, Gemini, plus various image/video/music generation models) under a single unified API gateway.

Getting Started

Platform Overview

With just one API key, access all top-tier models and enjoy unified billing, rate limiting, logging, and automatic failover. If you are already using OpenAI, Anthropic, or Google SDKs, simply change the base_url and api_key to seamlessly integrate.

Core Advantages

Unified API Interface

Fully compatible with native OpenAI, Claude, and Gemini protocols, allowing integration with zero code refactoring.

Minimalist Authentication & Billing

One account, one key, and a unified Quota pool that supports cross-provider model calls.

High Availability & Failover

Built-in enterprise-grade automatic failover and smart routing dispatch ensures seamless switching to backup nodes during single provider outages.

Full Modality Ecosystem

Not just for LLMs, deeply aggregates image, video, audio, and real-time voice generation and processing capabilities.

Architectural Highlights

Seamless Replacement & Multi-Protocol Support

Besides supporting OpenAI's `/v1/chat/completions`, we also support Claude's Messages API and Gemini's generateContent protocols. Fields are passed through as-is, requiring almost no code modifications.

Powerful Multimodal Task Processing

For long-running tasks like generating video or audio, we provide standard Submit & Poll async interface mechanisms to ensure high reliability in large-scale production environments.

Granular Rate Limiting & Risk Control

Supports setting independent request frequency limits, quota caps, and IP whitelists for different API keys, maximizing security isolation while ensuring business stability.