BYOM

Bring your own Model into OneRouter, OneRouter provides unified management for your Self-Deployed Models and public cloud MaaS APIs.

What is BYOM?

BYOM (Bring Your Own Model) is our flexible approach that lets you leverage your custom-built AI models within our ecosystem. Rather than being limited to public models(Claude 4.5、Gemini3. etc.), BYOM empowers you to host and integrate models you've developed and fine-tuned for your specific needs.

Introducing OneRouter

OneRouter is an unified management layer that makes BYOM seamless and powerful. It allows you to:

• Host your custom models on our platform while maintaining full ownership

• Integrate effortlessly with our existing infrastructure and services

• Manage all your models through a single, intuitive control plane

Key Features

Unified Scheduling/Orchestration Layer

OneRouter intelligently routes requests to the appropriate model based on your business logic and configuration preferences.

Smart Fallback Mechanism

Never worry about downtime. If your primary model encounters issues, OneRouter automatically redirects traffic to backup models, ensuring continuous service availability.

OneRouter offers a set of configurable metrics to dynamically determine when to fall back to alternative models and how to execute these fallback strategies,

including:

• Latency

• Time To First Token (TTFT)

• Tokens Per Minute (TPM)

• Requests Per Minute (RPM)

Unified Performance Monitoring

With OneRouter, you can monitor performance metrics for all public MaaS (Model-as-a-Service) APIs and your self-deployed models from a single, unified dashboard.

Why Choose BYOM with OneRouter?

Enhanced Privacy & Security

Keep your models in your own trusted environment. No need to upload your model to external platforms - your models stay where you control them, maximizing privacy and data security.

Superior Stability

Enjoy the best of both worlds: maintain security while leveraging OneRouter's intelligent fallback mechanisms to significantly improve overall service stability and reliability.

Perfect for Hybrid Deployments

In today's complex environments where organizations use both public Model-as-a-Service APIs and self-hosted custom models, OneRouter provides the ideal balance of:

• Operational stability

• Data privacy protection

• Cost efficiency

OneRouter is the optimal solution for organizations looking to maintain control of their AI assets while enhancing performance and reliability across their entire model ecosystem.

Quick Start

Users can register their self-deployed model endpoints on the OneRouter dashboard, allowing them to specify and call these models using the model parameter.

Models registered by users are available exclusively to the user who registered them. OneRouter implements permission isolation to ensure each user's custom models remain private and accessible only to them.