We are announcing a series of product launches and enhancements, expanding our web platform to build, scale, and secure AI-powered applications.
Why AI Inference on Azion?
Traditional cloud-delivered AI is often expensive, slow, and doesn’t always meet security, privacy, or compliance requirements. By moving inference to Azion’s Platform, we allow you to:
- Reduce latency dramatically by processing intelligence right next to users and devices.
- Lower costs by minimizing data transfer and expensive centralized resources.
- Stay resilient—your mission-critical AI apps keep working even if the connection to the cloud is lost.
- Comply with regulations because your data and models can remain in-country or within custom geographic zones.
Learn more about AI Inference:
- Learn more about the launch
- AI Inference page
- See what it solves for you
- Documentation
- Release notes
- Case study
Major New Features & Capabilities
AI Models with a Distributed Architecture
With Azion AI Inference, you can execute modern LLMs like Mistral, InternVL3, and Qwen on Azion’s distributed infrastructure. You can fine-tune the available models using Low-Rank Adaptation (LoRA) for parameter-efficient adaptation and cost-effective model customization.
OpenAI-API Compatibility: Rapid Migration, Simple Integration
Azion’s AI Inference is natively compatible with OpenAI’s API specification, so you can reuse existing prompts, workflows, and SDKs.
Benefit: Move your GenAI applications from the public cloud directly to the edge with minimal code changes; accelerate proofs-of-concept, hackathons, or enterprise migrations.
Distributed Vector Database for Real-Time Semantic Search
Azion AI Inference integrates deeply with Azion SQL Database—a globally replicated vector and relational database—allowing you to run hybrid search, semantic similarity, and advanced querying seamlessly across all locations.
- Instantly recommend content, products, or responses based on vector similarity, regardless of user location.
- Power AI-enabled “search as you type” UX for e-commerce or enterprise portals.
AI Agents, RAG Pipelines, and LangGraph Support
Go beyond basic chatbots:
- Build intelligent AI agents whose logic, memory, and tools execute in real time, calling Azion Functions or ingesting external APIs.
- Build RAG (Retrieval-Augmented Generation) pipelines: bring your own data sources, documentation, and context into every answer for next-level accuracy.
- Use tools like LangGraph for stateful conversational bots or orchestrate complex multi-step reasoning.
Always-On, Resilient AI Apps
Azion’s unique distributed architecture means your AI apps keep running—regardless of temporary cloud outages or regional failures. Models deploy close to users, ensuring low-latency results in critical workflows (security, finance, industrial IoT, etc.), and supporting business continuity.
Privacy & Regulatory Compliance
Data processed by AI Inference remains where you choose: in-country, in-region, or globally—with full auditing, encryption, and access controls.
Industries served: fintech, healthcare/medical, government, retail, legal, and every regulated segment that needs fast, local intelligence.
Model Fine-tuning with LoRA
Fine-tune the available models to your domain and deploy instantly to Azion’s distributed network.
- Stay ahead with the latest models AND keep your data private and valuable.
- No GPU clusters or ML ops experience necessary for scaling inference and deployment.
Discover Azion’s Solutions for Building and Scaling Applications
We introduced key enhancements to power your web applications directly at the data source, delivering the best user experiences through our compute technology, featuring a suite of products: Functions, SQL Database, and Object Storage.
Learn more about build and scale applications with Azion:
- Learn more about the launch
- See what it solves for you
- Guides on how to implement
- Release notes
- Pricing
Explore what’s new for build and scale:
Functions: New Limits, Broader Compatibility, and Faster Deployments
Functions was designed for creating and running web applications without worrying about infrastructure or unnecessary costs. Your code runs directly on Azion’s distributed infrastructure, with ultra-low latency and instant scalability.
What’s new:
- Runtime limits have been increased: now you can run functions for up to 5 minutes with application bundles up to 20 MB, supporting more complex workloads.
- Node.js and WinterTC API compatibility has been significantly extended, making it easier to port modern applications and reduce vendor lock-in.
- Framework support now covers Next.js, Vue, React, Angular, Gatsby, Astro, and more—allowing you to deploy your projects with the stack you prefer.
- Dozens of new, production-ready templates are available in the Console, accelerating the launch and iteration of websites, e-commerce stores, blogs, and other common use cases.
SQL Database: Enabling Seamless Data Integration and Migration
SQL Database is our serverless, fully ACID-compliant database, optimized for read-intensive and distributed workloads. Running across hundreds of Azion data centers, it delivers low latency and real-time responses.
What’s new:
- The new SQL Database Shell CLI simplifies migrations from PostgreSQL, MySQL, and other platforms, reducing friction and accelerating adoption.
- Expanded integrations with popular tools like Prisma and LangGraph, plus templates for Turso, Neon, Fauna, Drizzle ORM, MongoDB, and REST-compatible databases make onboarding and data operations effortless.
- Open connectivity empowers you to integrate with any system or external tool, creating a flexible foundation for building and scaling distributed data services.
Object Storage: Reduced Transfer Costs and Easier Adoption
Object Storage offers scalable object storage for large volumes of unstructured data, with no egress cost between Azion Object Storage and your Azion Edge applications.
What’s new:
- New program for free, customer-wide migrations from S3-compatible object storage from major cloud providers to Azion; no data size limits (see eligibility).
- Zero egress fees for transfers between Edge Storage and applications running on Azion, enabling high-volume projects without additional data transfer costs.
- New dashboards and detailed usage logs in the Console will be available for customers, providing greater transparency and management for your storage environment.
Azion Platform: Faster Delivery and Reliability at Scale
What’s new:
- Network and reliability enhancements with new HTTP/2 and HTTP/3 optimizations delivering faster, more modern application experiences.
- Global replication architecture now supports billions of workloads with high availability, further strengthening the resilience and scalability critical to enterprise-grade deployments.
With these launches and improvements, you can build and scale applications efficiently, using frameworks like Angular, Astro, Gatsby, React, Vue, and Next.js.
Find production-ready templates to accelerate your next project and explore what you can achieve with the Azion Web Platform.
Developer Experience: Empowering Builders at Every Step
Azion’s commitment to developers goes far beyond infrastructure-we’re reimagining the entire developer journey, making it faster, easier, and more productive to build, test, and scale applications at the edge.
Seamless App Delivery with Instant Domains
Every new application deployed on Azion now receives an instantly accessible, human-readable domain like https://myproject.azion.app. No more waiting for DNS propagation or complex configuration-your apps are ready to share, demo, and test from the moment they’re built. And soon, you’ll be able to add custom domains directly from the Azion Console, making branded deployments effortless.
API v4: Modern, Secure, and Built for Automation
Our brand-new API v4 is designed for today’s automation needs, with robust JWT/Bearer authentication, full activity history via GraphQL, and a powerful workload resource model that lets you orchestrate distributed edge applications as a single entity. The async, event-driven core ensures high performance and resilience for even the most demanding workflows.
API v4 is rolling out for new accounts on May 30, 2025.
Frameworks, Tooling, and Marketplace: Build Your Way
Whether you’re using Next.js, React, Vue, Angular, Gatsby, Astro, or another modern framework, Azion offers native support and optimized build pipelines. The new Azion Bundler v5.1 delivers builds up to twice as fast, while our open-source CLI, templates, and DevTools empower you to automate, test, and deploy with confidence. Explore the Azion Marketplace for ready-to-use integrations across security, analytics, AI, and more.
Open Standards, No Lock-In
Azion champions open source and open standards (including WinterCG APIs), ensuring you always have freedom and flexibility-no vendor lock-in, ever. Native Infrastructure as Code support (Terraform, CLI, and full API) means you can automate everything, from provisioning to deployment, at any scale.
Developer Resources
- Platform Overview
- CLI Reference
- API Reference
- Azion Lib
- DevTools
- Runtime Overview)
- Open Source Bundler
Azion’s platform is built for developers who demand speed, control, and innovation-so you can build, secure, and scale applications your way, everywhere in the world.
Getting Started
Azion is opening up a new frontier for building and scaling applications—enabling the boldest ideas, at the speed the world demands.
Ready to put all of this to work for your business?
Contact our experts, try our launch templates, or join our discord community.
