What Vercel shipped in March 2026

posts – Deployment Overview – Vercel

This content outlines key features of Vercel's deployment capabilities, focusing on three primary tools: Runtime Logs, Observability, and Web Analytics. Runtime Logs enable developers to view and debu…

Understand production from the inside out.

Vercel has enhanced its observability capabilities for frontend applications with real-time infrastructure and traffic insights, positioning itself as a 'mission control' for monitoring. Key updates i…

Build and deliver better web experiences. Faster.

Vercel has published a series of resources highlighting its focus on composable commerce, AI integration, and performance optimization for web development. The company emphasizes the advantages of com…

Create React App on Vercel: Enhanced Deployment and Observability Features

Vercel has published a technical guide outlining how to deploy and optimize Create React App (CRA) projects on its platform, emphasizing performance, collaboration, and observability features. The mos…

Backends on Vercel: Zero-Configuration Framework Support and Fluid Compute

Vercel now supports deploying a wide range of popular backend frameworks with zero-configuration, optimizing application performance through its infrastructure. Key benefits include Fluid Compute for …

Elysia on Vercel

Vercel now supports deploying Elysia, a TypeScript-first backend framework optimized for Bun, with zero configuration. The integration enables developers to build and deploy fast, type-safe backend se…

Koa on Vercel

Vercel now supports deploying Koa applications with zero configuration, enabling developers to run Koa-based web applications and APIs directly on Vercel’s platform. The integration leverages Vercel F…

NestJS on Vercel

Vercel now supports zero-configuration deployment of NestJS applications, enabling developers to deploy scalable backend services without manual setup. The integration leverages Vercel Functions, whic…

xmcp on Vercel

Vercel has introduced `xmcp`, a TypeScript-first framework for building MCP-compatible backends, enabling developers to deploy serverless applications with zero configuration. The framework provides a…

Scopes and Permissions Update for Sign in with Vercel

Vercel has updated its documentation for 'Sign in with Vercel' to clarify how scopes and permissions function within its authentication system. Scopes now explicitly define the data included in the ID…

Vercel Enhances Nuxt Integration with Simplified Deployment and Advanced Routing

Vercel has enhanced its integration with Nuxt, an open-source Vue framework, by providing streamlined deployment and configuration options for Nuxt applications on its platform. The most significant c…

Open Source Software

Vercel emphasizes its commitment to open source software by highlighting key projects it maintains or significantly contributes to, including Next.js, Turbo, AI SDK, Svelte, Nuxt, Nitro, Workflow DevK…

Axios package compromise and remediation steps

On March 31, 2026, the `axios` npm package was compromised in an active supply chain attack, with malicious versions (`axios@1.14.1`, `axios@0.30.4`, and `plain-crypto-js@4.2.1`) identified. Vercel, w…

Transfer Marketplace resources between teams

Vercel has introduced a new feature allowing users to transfer Marketplace resources between teams directly from the dashboard, eliminating the need for API-based transfers. This change simplifies res…

Supply Chain Attack on Axios Pulls Malicious Dependency from npm

A critical supply chain attack compromised the widely used JavaScript HTTP client Axios by injecting a malicious dependency, plain-crypto-js@4.2.1, into npm releases 1.14.1 and 0.30.4. The attack leve…

How FLORA shipped a creative agent on Vercel's AI stack

FLORA, a creative workflow platform, launched FAUNA, an AI-powered creative agent designed to streamline visual ideation for designers. Unlike FLORA’s previous node-based canvas, which required manual…

Vercel CDN now respects Cache-Control headers from external origins by default

Vercel is changing its CDN behavior to respect `Cache-Control` headers from external origins by default, starting April 6, 2026. Previously, responses proxied to external origins were uncached unless …

Agent responsibly: How Vercel is adapting to AI-driven development

Vercel’s latest blog post highlights the risks of blindly deploying AI-generated code, emphasizing that while coding agents can dramatically boost productivity, they also introduce new dangers when te…

Pricing discovered for Vercel

Pricing updated for Vercel: - New tier: Hobby ($0/mo) - New tier: Pro ($20/mo) - New tier: Enterprise (Custom pricing) - New tier: Observability Plus ($10/mo) - New tier: Web Analytics Plus ($10/mo) -…

Explore VercelEnterprise.

Vercel has introduced a free trial for its Vercel Enterprise offering, targeting organizations seeking scalable frontend development solutions. The trial provides access to enterprise-grade features, …

Vercel plugin now supported on OpenAI Codex and Codex CLI

Vercel has announced support for its plugin on OpenAI Codex and the Codex CLI, enabling teams to integrate over 39 platform skills, three specialist agents, and real-time code validation into their de…

Automatic persistence now in beta on Vercel Sandbox

Vercel has introduced automatic persistence in beta for its Sandbox feature, enabling sandboxes to save their filesystem state when stopped and restore it upon resumption without manual snapshots. Thi…

Vercel Sandboxes now allow unique, customizable names

Vercel has introduced a new feature for its Sandboxes product, allowing users to assign unique, customizable names to sandboxes instead of relying on ID-based identification. This change simplifies th…

MiniMax M2.7 High Speed

Vercel has introduced the MiniMax M2.7 High Speed model via its AI Gateway, offering the same performance as its predecessor but with significantly faster output speeds of approximately 100 tokens per…

Vercel launches Custom Reporting API for AI Gateway to unify AI spend data

Vercel has launched a Custom Reporting API for its AI Gateway in beta, targeting Pro and Enterprise plan teams. The API provides unified, programmatic access to cost, token usage, and request volume d…

Vercel Ship 26

Vercel has announced Vercel Ship 26, an upcoming event focused on enabling developers to build, scale, and secure intelligent applications and agents. The event emphasizes supporting teams at all stag…

Chat SDK now supports concurrent message handling

Vercel’s Chat SDK has introduced a new `concurrency` option for the `Chat` class, enabling developers to manage how incoming messages are processed when multiple messages arrive before prior ones fini…

WordPress 7.0 introduces built-in AI Client for generative AI integration

WordPress 7.0 adds a provider-agnostic AI Client API enabling plugins to send prompts to AI models and receive results through a consistent interface. The API supports text, image, speech, and video g…

Chat SDK now supports scheduled Slack messages

Vercel’s Chat SDK has introduced a new feature enabling scheduled Slack messages, allowing users to automate message delivery at future times. The update leverages the `thread.schedule()` method, wher…

Elastic build machines now available in beta

Vercel has introduced elastic build machines in beta for all paid plans, enabling teams to optimize build performance without manual project-level adjustments. The system automatically evaluates each …

Next.js at Scale

Vercel is hosting a live session titled 'Next.js at Scale' to discuss the challenges and strategies of self-hosting Next.js applications under high traffic. The session will cover caching behavior und…

Vercel acquires new.website to enhance v0 platform

Vercel has acquired new.website, a company focused on simplifying website creation with built-in tools like forms, databases, SEO features, and content management. The acquisition aims to accelerate V…

View specific error codes in runtime logs

Vercel has introduced a new feature that allows developers to view specific error codes directly within runtime logs in the Vercel dashboard. Previously, runtime logs only displayed HTTP status respon…

SERHANT.'s Playbook for Rapid AI Iteration with Vercel's AI SDK and AI Gateway

SERHANT., an AI-native real estate and media company, leveraged Vercel’s AI SDK and AI Gateway to rapidly iterate and scale its proprietary AI platform, S.MPLE, from an internal pilot to over 900 real…

Enterprise teams can now set their default build machine

Vercel has introduced a new feature allowing enterprise team owners to set a default build machine at the team level, which automatically applies to newly created projects. While existing projects ret…

Activity Log now available in Vercel CLI

Vercel has introduced the `vercel activity` command to its CLI, enabling users to query activity logs directly from the command line. The activity log tracks all team activities, including the user re…

Vercel acquires new.website to enhance v0’s AI-powered website-building capabilities

Vercel’s v0 has acquired new.website, a platform focused on simplifying website creation with built-in tools like forms, databases, and SEO features. The acquisition aims to enhance v0’s agent capabil…

Turborepo 2.9 delivers 91% performance boost via AI-assisted optimization

Turborepo, Vercel’s open-source build system for monorepos, has achieved a 91% performance improvement in its task graph computation, reducing Time to First Task from 8.1 seconds to 716 milliseconds i…

Sandbox SDK adds file permission control

Vercel’s Sandbox SDK version 1.9.0 introduces direct file permission control via the `writeFiles` API, allowing developers to set file permissions (e.g., executable scripts) in a single operation by p…

Chat SDK AMA: Build chat-native AI agents with one codebase

Vercel has introduced Chat SDK, a tool that allows developers to build chat-native AI agents with a single codebase, eliminating the need for separate integrations for each platform. This unified Type…

Vercel Fundamental Concepts

This page outlines Vercel’s core technical concepts, detailing how its platform processes requests, builds code, and manages compute resources. It explains the request lifecycle through Vercel’s globa…

Vercel details self-driving infrastructure and request lifecycle

Vercel has detailed its self-driving infrastructure, emphasizing how requests flow through its global network to deliver low-latency, scalable, and secure applications. The platform leverages a 100+ g…

Vercel Build System Documentation

Vercel’s documentation outlines its automated build system, which transforms source code into deployable assets for global distribution. When code is pushed to a connected Git repository, CLI, dashboa…

Vercel Introduces Fluid Compute: A Hybrid Model for Optimized Concurrency

Vercel has introduced Fluid Compute, a hybrid compute model that bridges traditional servers and serverless architectures to optimize performance, cost, and scalability. Unlike traditional serverless,…

Vercel AI Gateway Introduces OpenAI Service Tiers for Cost and Latency Control

Vercel’s AI Gateway now supports OpenAI’s service tiers, allowing developers to control processing priority, latency, and cost for OpenAI models. Three tiers are available: `default` (standard process…

Vercel Plugin for AI Coding Agents Launched

Vercel has launched a new plugin designed to transform AI coding agents into Vercel experts by embedding deep, context-aware knowledge of the entire Vercel ecosystem. The plugin pre-loads agents with …

Vercel Integrates Domain Connect for Automated DNS Configuration

Vercel has integrated Domain Connect, an open protocol enabling third-party services to automatically configure DNS records for domains managed on Vercel’s platform. This integration eliminates manual…

Vercel is now available in Stripe Projects

Vercel has integrated its platform into Stripe Projects, enabling users to deploy and manage Vercel projects directly from the terminal via a CLI-based workflow. This integration, available in develop…

Vercel launches Chat SDK to simplify AI agent deployment across chat platforms

Vercel has launched the Chat SDK, a TypeScript library designed to simplify the deployment of AI agents across multiple chat platforms from a single codebase. The core challenge addressed is the fragm…

Durable Migrates Entire Stack to Vercel for AI-Powered Business Platform

Durable, an AI-powered business builder platform, announced a strategic migration of its entire infrastructure to Vercel, enabling it to serve 3 million customers with just 6 engineers while processin…

MiniMax M2.7 is live on AI Gateway

MiniMax M2.7, a major upgrade to the MiniMax M2-series models, is now available on Vercel’s AI Gateway in two variants: standard and high-speed. The new model enhances software engineering, agentic wo…

Two startups at global scale without DevOps

Vercel highlights two AI-native startups, Leonardo.AI and Relevance AI, that operate at global scale without dedicated DevOps teams by leveraging Vercel’s platform. Leonardo.AI processes 4.5 million i…

Vercel: Durable leverages Vercel’s stack for AI-native multi-tenant platform

Durable, an AI business builder serving 3M customers with 6 engineers, migrated its entire stack to Vercel to support multi-tenant AI workloads. The move enabled 10x engineering leverage, 3-4x lower i…

v0 now includes diff view to review code changes

Vercel has introduced a diff view feature in v0, enabling users to review code changes directly within the interface. This update allows developers to see modifications file by file, including line ad…

Build knowledge agents without embeddings

Vercel has introduced a new approach to building knowledge agents by replacing traditional vector databases and embedding models with a filesystem-based architecture. The company observed that embeddi…

Use GPT-5.4 Mini and Nano on AI Gateway

Vercel has integrated OpenAI’s GPT-5.4 Mini and GPT-5.4 Nano models into its AI Gateway, expanding its AI model offerings for developers. These models are optimized for sub-agent workflows, where mult…

End-to-end encryption for Vercel Workflow

Vercel has introduced end-to-end encryption for its Workflow DevKit, automatically encrypting all user data without requiring code changes. This includes workflow inputs, step arguments, return values…

Vercel launches Agent: AI-powered code review, debugging, and installation tools

Vercel introduced Vercel Agent, an AI-powered suite of development tools including automated code reviews, anomaly investigation, and one-click integration of analytics and speed insights. Features le…

New GitHub App permissions for Actions and Workflows

Vercel has introduced new GitHub App permissions requiring repository access to Actions (read) and Workflows (read & write) for its GitHub App. These changes enable Vercel Agent to read CI failure log…

Introducing the Vercel plugin for coding agents

Vercel has launched a new plugin for coding agents (Claude Code and Cursor) that enhances their ability to understand and interact with Vercel projects by leveraging a real-time knowledge graph. The p…

Updates to Terms of Service - March 2026

Vercel has updated its Terms of Service and Privacy Policy to reflect changes in how it handles data to support agentic infrastructure and AI model training. The most significant change is the introdu…

Vercel GitHub App Updated Permissions

Vercel has updated the GitHub App to request two new repository permissions: Actions (read) and Workflows (read & write). These changes enable the Vercel Agent to access workflow run logs for diagnosi…

Vercel Open Source Program: Winter 2026 Cohort Announced

Vercel has announced its Winter 2026 cohort for the Vercel Open Source Program, expanding its initiative to support open source maintainers with resources, credits, and infrastructure. This marks the …

Vercel expands GitHub integration with automated deployments and preview URLs

Vercel for GitHub now automatically deploys every GitHub push, providing Preview Deployment URLs and Custom Domain updates. The integration supports GitHub Free, Team, Enterprise Cloud, and Enterprise…

Vercel Integrates OpenAI Codex with AI Gateway

Vercel has integrated OpenAI Codex with its AI Gateway, allowing users to route requests through multiple AI providers, monitor traffic, and view detailed traces. This setup enables users to configure…

Vercel Introduces Web Analytics

Vercel has introduced Web Analytics, a feature that provides detailed insights into website visitors. This tool tracks top pages, referrers, and demographics such as location, operating systems, and b…

Speed Insights Overview

Vercel's Speed Insights provides detailed performance metrics for websites, based on Core Web Vitals, enabling data-driven optimization decisions. The feature offers a dashboard with various views, in…

LiteLLM server now supported on Vercel

Vercel now supports deploying LiteLLM server, enabling developers to access large language models (LLMs) through an OpenAI-compatible gateway. This gateway connects to any supported provider, includin…

Meet the 2026 Vercel AI Accelerator Cohort

Vercel has announced the 2026 cohort of its AI Accelerator program, selecting 39 early-stage AI startups from the US, Europe, Asia, and Latin America to participate in a six-week intensive program. Th…

Streamdown 2.5 Released with KaTeX, Staggered Animations, and Code Block Improvements

Vercel’s Streamdown 2.5 introduces significant enhancements to its React component library for rendering streaming markdown in AI-powered applications. The update adds inline KaTeX support, which auto…

Vercel Introduces New `vercel api` Command in Beta

Vercel has introduced a new beta command called `vercel api` that allows users to make authenticated HTTP requests to the Vercel API directly from their terminal. This command automates authentication…

Vercel Introduces Activity Command in CLI

The Vercel CLI now includes an 'activity' command that allows users to view and filter activity events for their projects. This command can display events scoped to a specific project or across an ent…

Vercel launches Sandbox for safe, isolated code execution

Vercel introduced Sandbox, a compute primitive that runs untrusted or user-generated code in isolated, ephemeral Linux VMs using Firecracker microVMs. It supports Node.js and Python runtimes with fast…

Next-forge 6 is now available

Next-forge 6, a production-grade Turborepo template for Next.js apps, has been released with several developer experience (DX) improvements. The update introduces a next-forge skill for agents, making…

Vercel now supports Domain Connect as a DNS Provider

Vercel has integrated Domain Connect as a DNS provider, allowing external services to configure Vercel-hosted domains with one-click setup. This eliminates manual DNS record copying, reducing errors a…

AI Elements 1.9

Vercel's AI Elements 1.9 update introduces several new features and improvements. The update includes a new AI Elements skill that enhances agents' ability to build and use composable AI interfaces. A…

How Notion Workers run untrusted code at scale with Vercel Sandbox

Notion Workers, a feature allowing developers to extend Notion's platform with custom code, runs on Vercel Sandbox. This integration ensures secure execution of untrusted code by using Firecracker mic…

Deprecating the DHE cipher suite for TLS connections

Vercel will remove support for the legacy DHE-RSA-AES256-GCM-SHA384 cipher suite on June 30th, 2026. This change affects automated systems, security scanners, and HTTP clients with non-standard TLS co…

Provider Timeouts

Vercel has introduced provider timeouts for its AI Gateway, allowing users to set per-provider timeouts to trigger fast failover when a provider is slow to respond. This feature is particularly useful…

Chat SDK adds WhatsApp adapter support

Vercel's Chat SDK has added support for WhatsApp, extending its single-codebase approach to include Slack, Discord, GitHub, Teams, and Telegram. This new WhatsApp adapter enables teams to build bots t…

Vercel Flags are now optimized for agents

Vercel has enhanced its CLI to support programmatic flag management, allowing teams to create and manage feature flags directly from the terminal. This update eliminates the need to access the dashboa…

Improved data collection for Web Analytics and Speed Insights with resilient intake

Vercel has released version 2 of its Web Analytics and Speed Insights packages, introducing resilient intake to improve data collection reliability. This update dynamically discovers endpoints, ensuri…

Subscribe to webhook events for Vercel Flags

Vercel has introduced the ability to subscribe to webhook events for feature flags, providing deeper visibility into flag operations. This update allows teams to track the creation, modification, and …

@resend/chat-sdk-adapter

The Resend adapter for Vercel Chat SDK enables bidirectional email communication, allowing users to receive emails via Resend webhooks and send emails through the Resend API. This integration supports…

@beeper/chat-adapter-matrix

The Matrix adapter for Chat SDK has been updated to run over Matrix sync instead of webhooks, enabling integration with Beeper conversations and bridged networks like WhatsApp, Telegram, Instagram, an…

Community Providers for AI SDK

The AI SDK now supports a variety of community providers that adhere to the Language Model Specification. This allows developers to write custom providers compatible with the AI SDK, expanding the eco…

chat-adapter-imessage

The Chat SDK has introduced an iMessage community adapter that supports both local and remote integration with iMessage. The remote mode, recommended for production, connects to Photon's managed iMess…

Vercel Sandbox now supports 1 vCPU + 2 GB RAM configurations

Vercel has updated its Sandbox offering to support configurations with 1 vCPU and 2 GB of RAM, in addition to the existing default of 2 vCPUs and 4 GB of RAM. This change is aimed at users with single…

How we run Vercel's CDN in front of Discourse

Vercel has implemented its CDN to front its Discourse-based community platform, enhancing security and functionality without migrating the entire application. This setup provides DDoS protection, web …

Chat SDK now has an adapter directory

Vercel's Chat SDK now features an adapter directory, allowing users to search for platform and state adapters created by Vercel, vendors, and the community. Official adapters are maintained by the cor…

Vercel introduces Deployment Checks to decouple builds from releases

Vercel added Deployment Checks to enforce conditions before promoting production builds to custom domains, decoupling builds from releases for safer deployments. Teams can now require GitHub Actions o…

Chat SDK adds PostgreSQL state adapter

Vercel's Chat SDK has added support for PostgreSQL as a state backend, joining Redis and ioredis. This update allows teams already using PostgreSQL to manage subscriptions, distributed locks, and key-…

Vite + Nitro on Vercel

Vercel has introduced support for Nitro, a universal server toolkit, to enhance Vite applications with server-side rendering (SSR), API routes, and server middleware. Nitro, which powers frameworks li…

Vercel Introduces Stripe Billing Integration for AI Gateway

Vercel has introduced a new feature that allows users to bill customers for AI usage through AI Gateway by integrating with Stripe's metered billing. This integration enables automatic emission of met…

CNBC: A Fortt Knox Conversation with Guillermo Rauch

9 Mar 2026 Perspectives ↗ CNBC: A Fortt Knox Conversation with Guillermo Rauch Vercel CEO Guillermo Rauch describes an unprecedented wave of “shadow IT” as enterprise employees outpace their own IT te…

Projects and Deployments on Vercel

Vercel's documentation outlines the concepts of projects and deployments within their platform. A project is defined as an application deployed to Vercel, which can be connected to a single repository…

Provider Filtering & Ordering in Vercel's AI Gateway

Vercel's AI Gateway now allows users to control provider selection and ordering for AI model requests. By default, the gateway selects providers based on uptime and latency, but users can override thi…

Vercel AI Gateway Adds Reasoning Capabilities

Vercel's AI Gateway now supports reasoning capabilities across multiple AI providers, including OpenAI, Anthropic, Google, Vertex AI, and Amazon Bedrock. This feature allows AI models to 'think' befor…

Vercel Introduces OpenAI Reasoning Models

Vercel has introduced OpenAI reasoning models that allow users to control the depth of reasoning and receive summaries of the model's thought process. The supported models include various versions of …

Vercel Introduces Adaptive Thinking in Anthropic Claude 4.6 Models

Vercel has introduced adaptive thinking in its Anthropic Claude 4.6 models, allowing dynamic decision-making on reasoning effort levels. This feature is available in both Claude Opus 4.6 and Claude So…

Vercel Enhances AI Gateway with Google's Gemini Models

Vercel has enhanced its AI Gateway capabilities by integrating Google's Gemini 2.5, 3, and 3.1 series models, which feature an internal 'thinking process' to improve reasoning and multi-step planning.…

Vercel Adds Amazon Bedrock Reasoning Support

Vercel has introduced support for Amazon Bedrock's reasoning features for Anthropic models. The configuration varies based on the model used. For Claude 4.6, adaptive reasoning is supported with a 'ma…

Vercel Introduces AI SDK for AI-Powered TypeScript Applications

Vercel has introduced the AI SDK, a recommended tool for building AI-powered TypeScript applications with AI Gateway. The SDK allows developers to pass model strings directly to functions, automatical…

OpenAI Responses API

The OpenAI Responses API on Vercel is introduced as a modern alternative to the Chat Completions API. It allows users to route requests to various AI providers like OpenAI, Anthropic, and Google by us…

Vercel Introduces OpenAI Chat Completions API

Vercel has introduced the OpenAI Chat Completions API through its AI Gateway, allowing users to access multiple AI providers via a familiar interface. This API enables developers to use existing OpenA…

Vercel Introduces Automatic Caching for AI Gateway

Vercel has introduced automatic caching for its AI Gateway, which simplifies prompt caching for developers. The feature automatically handles caching strategies based on the provider, with some provid…

Vercel Introduces Domain Purchasing Feature

Vercel has introduced a new feature allowing users to purchase domains directly through their platform. This service enables users to find, select, and buy domains that align with their brand, with WH…

Use an existing domain

Vercel has provided a tutorial on how to integrate an existing domain with their platform. This feature allows users to maintain their established online presence while leveraging Vercel's powerful fe…

Vercel Enhances Collaboration Features with Automatic Deployments and Comments

Vercel has enhanced its collaboration features to streamline team workflows. The platform now automatically deploys every commit pushed to a Git repository, creating preview environments for non-main …

The article outlines the fundamental concepts of Vercel, focusing on how requests are processed through its global network, including load balancing, security, caching, and compute. It also explains t

How requests flow through Vercel

Vercel's infrastructure is designed to route, secure, and serve user requests efficiently through a global network of servers. When a user requests a site hosted on Vercel, their browser performs a DN…

Getting started with Vercel

Vercel is an end-to-end platform for developers that simplifies the process of building and deploying web apps. It supports popular frontend frameworks out-of-the-box and offers a globally distributed…

Agentic Commerce in 2026: What’s Actually Working

Vercel is hosting an event focused on Agentic Commerce in 2026, featuring speakers from Stripe, Elkjøp Nordic, Orium, and Vercel. The event aims to provide insights into what's currently delivering RO…

Home - Why Engineering Velocity Still Breaks Morning Briefing Sydney

The content provided is a summary of an event titled 'Why Engineering Velocity Still Breaks Sydney'. The event focuses on discussing the impact of engineering velocity on Sydney, likely in the context…

Chat SDK adds table rendering and streaming markdown

Vercel's Chat SDK has introduced native table rendering and improved streaming markdown capabilities. The new Table() component allows developers to render tables across all platform adapters, with ea…

Deployment step now 15% faster

Vercel has improved its deployment step, making it 15% faster on average, with more complex projects seeing up to a 3.7-second reduction in deployment time. This optimization was achieved by improving…

v0 API now supports custom MCP servers

Vercel's v0 API now supports connecting to any custom MCP server, allowing teams to configure new servers programmatically by providing endpoint and authentication details. This feature enables users …

AI Gateway supports OpenAI's Responses API

Vercel's AI Gateway now supports OpenAI's Responses API, offering a modern alternative to the Chat Completions API. This update allows users to route requests through AI Gateway using the Responses AP…

Skip unaffected builds for projects in Bun monorepos

Vercel has extended its ability to skip unaffected builds in monorepos to support Bun lockfiles, a feature already available for other package managers. This update allows teams using Bun to reduce un…

AI Gateway now supports OpenAI's Responses API

Vercel's AI Gateway now supports OpenAI's Responses API, allowing users to route requests to models from various providers through a single interface. This update introduces a flatter input/output for…

Create private blob stores with a single click in v0

Vercel has introduced a new feature in v0 that allows teams to create private and public blob stores with a single click. This update enables users to select their preferred region and access type whe…

Stripe is now generally available on the Marketplace and v0

Vercel has announced the general availability of Stripe integration on its Marketplace. This update allows users to connect their production Stripe accounts to Vercel, enabling real payment processing…

Customize timeouts for faster automatic failover on Vercel AI Gateway

Vercel AI Gateway now supports custom timeouts for faster failover on AI Gateway. This feature allows users to set per-inference timeouts for providers, enabling quicker fallbacks if a provider does n…

Provider-level custom timeouts available on AI Gateway

Vercel's AI Gateway now supports provider-level custom timeouts, allowing users to set specific timeout durations for each provider. This feature enables faster failover if a provider does not respond…

Vercel's CDN now supports updating routing rules without a new deployment

Vercel has introduced a new feature allowing users to update routing rules within a project without requiring a new deployment. This update enables users to set response headers or rewrites to externa…

GPT 5.4 is now on AI Gateway

Vercel has announced the availability of GPT-5.4 and GPT-5.4 Pro on its AI Gateway. This update brings significant improvements in agentic and reasoning capabilities, extending beyond coding to includ…

Streamdown 2.4: More customization, accessibility and custom rendering

Streamdown v2.4 introduces several new features and improvements for developers rendering markdown. The update allows teams to customize the appearance of their markdown output by overriding built-in …

Vercel's CDN gets a new dashboard experience

Vercel has introduced a new dashboard experience for its CDN, consolidating global traffic distribution tracking, top-line metrics, caching management, and routing rules into a single interface. The u…

From idea to secure checkout in minutes with Stripe

Vercel has announced the general availability of Stripe integration on its Marketplace and in v0, enabling developers to connect live Stripe accounts to their projects seamlessly. This update eliminat…

Run cron jobs from deployment summary

Vercel has introduced a new feature that allows users to run cron jobs directly from the deployment summary section of their deployments dashboard. This update simplifies the process of setting up and…

Vercel expands Deployment Protection with new scopes and methods

Vercel introduced Deployment Protection to restrict access to preview and production URLs, now offering three protection methods (Vercel Authentication, Password Protection, Trusted IPs) and multiple …

MCP Apps support on Vercel

Vercel has introduced support for MCP Apps, allowing teams to build and deploy these apps with full Next.js compatibility. MCP Apps are an open standard for embedded UIs that run inside iframes and co…

Vercel Workflow is now twice as fast

Vercel has significantly improved the performance of its Vercel Workflow platform, which is built on the open-source Workflow Development Kit (WDK). The server-side performance has doubled, with a 54%…

Vercel Sandbox now accepts environment variables at creation

Vercel has updated its Sandbox SDK and CLI to support setting environment variables at sandbox creation. This change allows environment variables to be automatically available to every command within …

Gemini 3.1 Flash Lite is now on AI Gateway

Vercel has announced the availability of Google's Gemini 3.1 Flash Lite model on its AI Gateway. This model surpasses its predecessor, 2.5 Flash Lite, in overall quality, particularly in translation, …

GPT 5.3 Chat is now on AI Gateway

Vercel has added GPT-5.3 Chat to its AI Gateway, enhancing tone, relevance, and conversational flow. The update reduces hallucination rates and improves response accuracy. Users can access this model …

Building Slack agents can be easy

Vercel has introduced a new Slack agent skill that simplifies the process of building and deploying Slack agents. This tool automates the coordination of various systems required to create a Slack age…

Scaling redirects to infinity on Vercel

Vercel has optimized its redirect system to handle millions of redirects efficiently. Previously, redirects were managed through routing rules and middleware, which became inefficient at scale. The ne…

Search code, repositories, users, issues, pull requests...

The conversation revolves around a GitHub pull request for a chat application. The main change involves migrating from Postgres to pg (node-postgres) as the state store, which is considered more suita…

Vercel CLI for Marketplace Integrations Optimized for Agents

Vercel has updated its CLI to optimize for AI agents, enabling them to autonomously discover, install, and retrieve setup instructions for Marketplace integrations. The new `discover` and `guide` comm…

Advancing Python typing

Vercel has announced a year-long research effort aimed at enhancing Python's type system, making it more expressive and composable, similar to TypeScript's programmable types but tailored for Python's…

Vercel integrates TanStack Start via Nitro for full-stack deployments

Vercel now supports deploying TanStack Start applications using the Nitro Vite plugin, enabling full-document SSR, streaming, server functions, and automatic scaling via Vercel Functions and Fluid Com…