Vercel
You can now access anomaly alerts and their details directly through the Vercel CLI using the vercel alerts command, allowing you to view alert information and AI investigation results without leaving the terminal.
You can now access anomaly alerts and their details directly through the Vercel CLI using the vercel alerts command, allowing you to view alert information and AI investigation results without leaving the terminal.
The Chat SDK now includes support for Messenger as a chat adapter, enabling features like messages, reactions, multimedia downloads, and more, with automatic fetching of display names from user profiles.
The Chat SDK now supports a web adapter, enabling the creation of browser-based chat UIs for various applications like in-product assistants and support agents. Developers can utilize the @ai-sdk/react useChat hook to stream replies live to the browser.
The Chat SDK now includes support for cross-platform conversation history, allowing users to maintain their message history across different platforms with new methods for managing transcripts.
A coordinated security release for Next.js has been issued, addressing 13 advisories related to various vulnerabilities including denial of service, middleware and proxy bypass, server-side request forgery, cache poisoning, and cross-site scripting. Users are urged to upgrade to the patched versions of React and Next.js immediately.
Pro teams can now select how Git committers to private repositories are added to their Vercel team, with options for auto approval or manual approval of new members.
Vercel now allows users to secure marketplace credentials by restricting access to production environments only, enhancing the protection of sensitive information. This change removes non-production access and requires Owner permissions to revert.
The Vercel CLI now allows users to query Observability Plus metrics using the new 'vercel metrics' command, enhancing the ability to analyze application performance and debug issues. This feature is currently in public beta for all teams with Observability Plus.
KIKO Milano improved their Black Friday operations by migrating to Vercel, eliminating manual infrastructure scaling and reducing app build times, allowing for multiple daily deployments and a focus on enhancing the ecommerce experience.
General Intelligence has successfully built an agent platform on Vercel, enabling founders to run companies entirely with AI agents, while automating 90% of their SRE work and managing complex infrastructure needs through Vercel's capabilities.
Vercel Sandbox now supports connections to hosted Postgres databases through its firewall by adjusting for the Postgres TLS negotiation flow, allowing users to connect without code or configuration changes. TLS is required for these connections, and specific encryption modes are not supported.
Grok 4.3 is now available on Vercel AI Gateway, featuring a 1M token context window and enhancements in accuracy, tool calling, and instruction following. Users can access it by setting the model to xai/grok-4.3 in the AI SDK.
Vercel introduces custom tags in beta for organizing and managing sandboxes, allowing teams to filter and track usage effectively across various environments and workflows.
Vercel now allows users to sign up for or upgrade to a Pro plan directly from Stripe Projects using shared payment tokens, enabling programmatic management of plan changes via the Stripe CLI.
Vercel now supports Native Deployment Checks, allowing teams to run lint and typecheck scripts in parallel with deployments, enhancing the deployment process with optional required checks and automated failure investigation.
The 2026 Vercel AI Accelerator culminated in a Demo Day where 39 teams showcased their AI applications after six weeks of intensive workshops and mentorship, with significant support from Vercel and partners. Each team received over $200K in resources to develop their projects.
Starting April 29th, Hobby plans will have a maximum deployment retention policy of 30 days, automatically removing older deployments while preserving the 10 most recent production and aliased deployments. Pro and Enterprise plans remain unaffected.
Vercel has introduced GPT-5.5 and GPT-5.5 Pro on the AI Gateway, enhancing capabilities for long-running tasks and improving token efficiency, with a unified API for model usage and performance optimizations.
DeepSeek V4 is now available on Vercel AI Gateway with two model variants, Pro and Flash, offering enhanced capabilities for coding and reasoning tasks. The AI Gateway also introduces a unified API for improved model management and performance optimizations.
Vercel has introduced GPT Image 2 on the AI Gateway, which enhances image generation capabilities with detailed instruction following and supports various visual styles at up to 2K resolution. Users can access this model through the AI SDK or model playground.
Kimi K2.6 from Moonshot AI is now available on Vercel AI Gateway, enhancing capabilities for long-horizon coding tasks across multiple programming languages and improving API interpretation and stability for autonomous agents.
Vercel's deployment retention policies now ensure that the latest preview deployment for branches with open or unmerged pull requests is preserved, allowing for shorter retention windows without the risk of losing active deployments. This change applies to all plans.
Zo Computer significantly improved AI reliability on Vercel, achieving a 20x reduction in retry rates and a 99.93% chat success rate, while also streamlining model integration and deployment processes.
Vercel Flags, a feature flag provider integrated into the Vercel platform, is now generally available, allowing users to create and manage feature flags with various targeting options directly in the Vercel Dashboard.
Claude Opus 4.7 is now available on Vercel AI Gateway, featuring enhancements for long-running tasks, improved memory management, and the introduction of task budgets. Users can set the model to anthropic/claude-opus-4.7 and utilize a new effort level 'xhigh'.
GitBook successfully migrated to Vercel, enabling sub-second content updates for 30,000 documentation sites by utilizing cache directives and improving cache invalidation processes, ensuring that only relevant sites are updated when changes are made.
Vercel has launched Workflows, a new programming model for durable execution that simplifies the management of long-running systems by integrating orchestration directly into application code, now generally available after a successful beta period.
Vercel has introduced Bytedance's Seedance 2.0 video generation model on AI Gateway, offering two variants (Standard and Fast) that support various input types and features like synchronized audio and video editing.
Vercel is reducing the pricing for Turbo build machines by 16%, with new rates effective from April 27, 2023. The updated pricing includes $0.105 per minute for Turbo machines, while Enhanced and Standard machines maintain their current rates.
Vercel now supports workflow run log filtering, allowing users to easily view and filter logs associated with specific workflow runs and steps directly from the Logs tab.