nRouter Expands Gateway to 100+ Multimodal Models

SUNNYVALE, CA, September 16, 2026 — nRouter, the enterprise AI gateway built by CloudAct Inc., today announced a major expansion of its catalog to support over 100 verified foundation models across four key modalities: Text, Audio, Image, and Video, accessible through a single unified endpoint.

As software organizations transition from isolated conversational chatbots to rich multimodal experiences, engineering teams face significant architectural overhead maintaining disparate SDKs, differing authentication paradigms, and vendor-specific payload formats. nRouter eliminates this friction by standardizing multimodal inference into a cohesive, high-performance API interface.

Universal Multi-Modality

nRouter now provides drop-in compatibility across the entire foundation model spectrum:

• Text & Frontier Reasoning: Seamless access to Claude 3.7 / 3.5 Sonnet, GPT-4.5 / o3, Gemini 2.5 Pro / Flash, DeepSeek V3 / R1, and Meta Llama 3.3.
• Vision & Document Analysis: Native image understanding and multimodal PDF extraction across Vertex AI, OpenAI, and Anthropic.
• Audio & Speech: High-accuracy speech-to-text transcription and expressive text-to-speech synthesis via OpenAI Whisper, ElevenLabs, and Vertex Audio.
• Image & Video Generation: Programmatic image generation via Imagen 3 and Flux, alongside cutting-edge video generation pipelines powered by Google Veo.

Intelligent Auto-Routing (nrouter/auto)

To help engineering teams maximize model performance while controlling runaway compute expenses, nRouter features algorithmic auto-routing. By passing model: "nrouter/auto", the gateway inspects incoming prompt token lengths, intent complexity, and real-time provider health:

• Routine Queries: Automatically dispatched to economical, high-speed models (such as Claude 3.5 Haiku, GPT-4o mini, or Gemini Flash), cutting token spend by up to 60%.
• Complex Multi-Step Reasoning: Intelligently escalated to top-tier frontier models when mathematical, architectural, or code-synthesis reasoning is required.
• Zero-Downtime Failover: If an upstream provider returns a 500 error or rate-limit ceiling, nRouter seamlessly retries on equivalent backup models without interrupting user sessions.

Seamless Developer Experience

Developers can integrate the full multimodal model fleet with a single configuration line using standard OpenAI or Anthropic SDKs by pointing the base URL to https://api.nrouter.ai/v1.

"Multi-model optionality is no longer a luxury for enterprise engineering teams — it is a mission-critical reliability requirement," said Rama Surasani, Founder and CEO of nRouter. "By expanding our gateway to over 100 verified models across text, vision, audio, and video, we give developers the superpower to adopt the best model for every task with zero proprietary lock-in."

Explore the Model Catalog

The live catalog of supported foundation models and real-time latency statistics can be browsed at nrouter.ai/models.

For media inquiries: sales@nrouter.ai. For developer support: sales@nrouter.ai.