ZenMux

ZenMux is an AI model gateway that gives developers one account and one API for a broad catalog of models. It speaks the OpenAI, Anthropic, and Google Vertex AI protocols, so existing clients can point at ZenMux with minimal changes, while a browser studio covers chat, image, and video generation. The service also adds routing, failover, usage analytics, and automatic compensation when a response is judged to have failed.

Key Features

• One API key and one billing account across many models, without separate provider accounts.
• Compatibility with the OpenAI, Anthropic, and Google Vertex AI request formats.
• Automatic failover across provider channels when a model is rate limited or unavailable.
• Auto routing that picks a model for a prompt based on quality and cost.
• Token-level usage dashboards for requests, spend, and model mix.
• AI insurance that compensates for hallucinated, slow, or low-throughput responses.
• A Studio interface for chat, image generation, and video generation.

Who is it for

• Developers building products that call several model providers.
• Teams moving from prototypes to production that need failover and cost controls.
• Studios and creators who want one interface for chat, images, and video.
• Platform engineers who need usage reporting per request and per token.

What can you do with ZenMux

• Point an OpenAI-compatible SDK at the gateway and switch between models.
• Compare model output and cost before committing a workload.
• Route production traffic with a fallback when a provider fails.
• Track token spend by model, key, and project.
• Generate images and video from the Studio.
• Claim compensation for requests that fail quality or latency checks.

How does ZenMux work

1. Create a ZenMux account and generate an API key.
2. Choose a subscription plan or top up a pay-as-you-go balance.
3. Set the gateway base URL in the SDK or client you already use.
4. Pick a model, or enable automatic routing for the request.
5. Monitor usage and compensation in the dashboard, then adjust limits.

Pricing

• Builder Plan: fixed monthly fee available at $0, $20, $100, and $200 tiers, with a weekly quota cap and 10-15 RPM.
• Pay As You Go: usage-based billing with token-level charges and no rate limit.
• Pay-as-you-go top-ups add 10% bonus credits.
• The pricing page describes the subscription plan as suited to development and prototyping, and pay-as-you-go as suited to production.

FAQ

Does ZenMux require separate accounts with each model provider?

No. You create one ZenMux account, generate a key, and call the models through the gateway, which sources access from official providers or authorized cloud partners.

Which API formats does ZenMux support?

The gateway is compatible with the OpenAI, Anthropic, and Google Vertex AI protocols, so existing SDKs and tools can often switch by changing the base URL.

What is AI insurance?

ZenMux monitors responses for hallucination signals, excessive latency, and low throughput, and automatically compensates eligible failures according to the plan terms.

Can I use the subscription plan in production?

The pricing page describes the fixed-fee Builder plan as intended for development, learning, and prototyping, and directs production workloads to pay-as-you-go billing.

What happens if a provider goes down?

ZenMux keeps backup channels for models and fails over automatically when a provider is rate limited, fails, or is unavailable in a region.