Amazon Nova Models: Now Available on Heroku

Building intelligent applications requires powerful, cost-effective AI. Today, we’re simplifying that process by making Amazon’s cutting-edge Nova models directly available via Heroku Managed Inference and Agents. Provisioning these models is as simple as attaching the add-on to your Heroku application, providing a direct, managed path for developers and businesses to leverage a new class of powerful and cost-effective AI models with unparalleled simplicity.

Heroku AI: balancing performance and cost with Nova models

The Amazon Nova family of models is engineered to provide an exceptional balance of performance and cost. Both nova-pro and nova-lite are optimized for modern development patterns like Retrieval-Augmented Generation (RAG) and building powerful AI agents.

While these models have powerful multimodal capabilities, our initial integration on Heroku focuses on their robust text-to-text functionality. Support for multimodal capabilities will be introduced in a future release.

Model comparison at a glance

nova-pro: for demanding, high-accuracy tasks

The nova-pro model is optimized for complex scenarios where accuracy and deep reasoning are critical. It excels at processing and reasoning over vast amounts of information, making it exceptionally well-suited for tasks that require a deep understanding of extensive context.

Use cases:

• Financial document analysis: Analyze the full scope of lengthy financial reports or contracts in a single pass.
• Sophisticated agentic workflows: Build autonomous agents that can maintain context across many steps and tool interactions.

nova-lite: for high-throughput, real-time applications

The nova-lite model is built for speed. It delivers rapid, low-cost inference, making it the ideal choice for high-volume applications that require immediate responsiveness, while still handling significant contextual information.

Use cases:

• Interactive customer support: Power intelligent chatbots that can recall long conversation histories to provide personalized support.
• Live document processing: Instantly classify and extract data from large documents as they are uploaded.

From text generation to action: powerful tool use with Nova and Heroku

What truly sets these models apart is their powerful ability to move beyond text and take action. Both nova-pro and nova-lite have tool use (also known as function calling) capability. This feature allows the models to go beyond generating text and interact with external systems.

Simplified by Heroku Managed Inference and Agents

We’ve made connecting these models to your apps seamless. Just attach the Heroku Managed Inference and Agents add-on, and you are ready to go. Heroku simplifies the operation with these models by providing two powerful, purpose-built endpoints:

An OpenAI-compatible endpoint: For straightforward chat and text-generation tasks, use the familiar /v1/chat/completions endpoint. This allows for easy integration with existing codebases and libraries designed for the OpenAI API.

A powerful agents endpoint: To build more sophisticated agentic workflows, the /v1/agents/heroku endpoint provides native support for both built-in Heroku tools (like running code or querying a database) and custom tools via the Model Context Protocol (MCP). This makes it simple to create AI agents that can take action within your application ecosystem.

This dual-endpoint approach gives you the flexibility to choose the right level of abstraction for your needs, from simple text generation to complex, multi-step agentic tasks.

Getting started

You can provision Nova models in both the US and EU regions. To attach a model to your application, simply attach the Heroku Managed inference and Agents addon or use the Heroku CLI. For example, to provision the nova-pro model:

heroku addons:create heroku-inference:nova-pro -a your-app-name

From there, your application is ready to start making calls to the model and building the next generation of intelligent, responsive applications.