Vercel

An open-source TypeScript framework for building AI-powered applications. The AI SDK provides a single, consistent interface to every major AI provider, so the applications we build for you can use Claude for analysis, ChatGPT for content, and Gemini for translation, all within the same system. Version 6 introduced agents as a first-class building block, with full MCP support and tool execution approval for safe, controlled automation.

• Provider-agnostic: connect to Anthropic, OpenAI, Google, and others through one interface
• Agents as first-class abstractions, define once and deploy across your applications
• Full MCP support for connecting AI to your enterprise tools and data sources
• Tool execution approval so humans stay in the loop on sensitive actions
• Built-in streaming for responsive, real-time AI interfaces

A single API that routes AI requests across multiple providers, with automatic failover, cost tracking, and performance monitoring built in. AI Gateway means your AI applications stay up even when a provider goes down, and you get unified visibility into usage, costs, and performance across every AI model you use.

• Sub-20ms latency routing across AI providers at the edge
• Automatic failover routes around provider outages without code changes
• Zero markup on AI tokens, including Bring Your Own Key support
• Unified observability dashboard for usage, costs, and performance across all providers
• Intelligent routing to select the best provider for each request

Isolated cloud environments built on Firecracker microVMs for running AI-generated code and agent workflows safely. Each sandbox gets its own filesystem, network, and process space, completely separated from your production infrastructure. When AI agents need to execute code, analyse data, or run automated workflows, Sandboxes keep everything contained.

• Firecracker microVM isolation, each sandbox is completely separated from infrastructure
• On-demand Linux environments with sudo access and package managers
• Ephemeral by design, sandboxes shut down automatically after use
• Up to 5 hours of execution time on Pro and Enterprise plans
• Open-source SDK and CLI for custom integration

An AI-powered development agent that builds full-stack applications from natural language descriptions. v0 can import existing code repositories, connect to databases, and generate production-ready applications with proper security controls and git workflows. For organisations building internal AI tools, v0 accelerates development from weeks to days.

• Full-stack application generation from natural language descriptions
• Imports existing GitHub repositories and environment configurations
• Enterprise security: SAML SSO, role-based access control, and audit logs
• Database integrations with Snowflake, AWS, and other enterprise data sources
• Generates production-ready code with proper git workflows

Vercel's compute model that combines the efficiency of serverless with the flexibility of traditional servers. Fluid Compute prioritises existing resources before creating new instances, meaning your AI applications scale smoothly under load without the cold-start delays that plague traditional serverless. For enterprise clients, multi-region failover is enabled by default.

• Scales functions before instances for faster, more efficient handling of AI workloads
• No hard scaling limits, handles tens of thousands of concurrent requests
• Multi-region failover enabled by default for enterprise
• Dense global compute runs workloads close to where your data lives
• Cost-efficient, you pay for active compute, not idle resources

The deployment and hosting layer that ties everything together. Vercel's Enterprise plan provides the security, compliance, and operational controls that regulated industries require, from SOC 2 Type 2 and HIPAA to a Web Application Firewall and dedicated compute environments. Your AI applications get enterprise-grade infrastructure without your team managing servers.

• SOC 2 Type 2 certified with annual third-party audits
• HIPAA compliance available for healthcare and regulated industries
• PCI DSS v4.0 attestation for payment-related applications
• Web Application Firewall for edge-based traffic filtering and monitoring
• GDPR compliant with configurable data retention policies