Vercel 在 2026年8月 发布的内容

Pricing change detected for Vercel

Pricing updated for Vercel: - Hobby: now includes 100 GB bandwidth / month, 10 GB storage / month, 100 GB data transfer / month, 10K function invocations / month, 10K edge requests / month - Pro: now …

Vercel releases design.md and public stylesheet for agent-generated pages

Vercel introduced design.md, a public file providing structured design guidance for AI agents to generate on-brand pages, and a public stylesheet (vercel-brand.css) to standardize visual elements. The…

Vercel Sandbox snapshot storage now billed daily instead of monthly

Vercel changed how Sandbox snapshot storage costs are calculated, switching from a monthly average to daily average usage to determine charges. The $0.08 per GB-month rate remains unchanged, but billi…

Vercel adds fx coding agent to AI SDK harness layer via new adapter

Vercel integrated its fx lightweight coding agent into the AI SDK harness layer by releasing the `@ai-sdk/harness-fx` adapter, enabling developers to run fx within applications using a single API. The…

Vercel adds per-user budgets to AI Gateway for granular spend control

Vercel introduced per-user spending budgets in AI Gateway, allowing teams to set monthly, weekly, or daily dollar limits for individual users. Budgets reset automatically and trigger email alerts at 5…

Vercel slashes MiniMax H3 and H3 Max prices by 50% on AI Gateway

Vercel introduced a 50% discount on MiniMax H3 and H3 Max models via AI Gateway from August 30 to September 13, 2026. The discount applies to all supported request durations and aspect ratios, with no…

Pricing updated for Vercel: - Pro: promotion changed — $20 included credit to use across resources - Enterprise: promotion changed — Contact sales team

Vercel enables no-code agent building and deployment for eve agents

Vercel introduced a new dashboard feature enabling users to build and deploy eve agents in a few clicks, automatically creating a Git repository and Vercel project. The builder supports defining agent…

Vercel CLI expands DNS, domains, and project commands

Vercel CLI version 59.6.2 introduces new commands for managing DNS records, domain renewals, project state, and team access directly from the terminal. These updates mirror dashboard and API capabilit…

Vercel adds Tencent’s Hy4 Preview Mixture-of-Experts model to AI Gateway

Vercel’s AI Gateway now supports Tencent’s Hy4 Preview, an open-source 770B-parameter Mixture-of-Experts model with 1M-token context windows, optimized for coding, document analysis, and scientific re…

Vercel launches Ling 3.0 Flash Fin AI model on AI Gateway with free access

Vercel added Ling 3.0 Flash Fin, a finance-focused AI model with a 256K-token context window and 32K output tokens, to its AI Gateway for free until September 25, 2026. Users can choose between a bill…

Vercel launches Workflow SDK: turning TypeScript into a durable execution engine

Vercel introduced Workflow SDK, a framework that turns TypeScript into a durable execution engine without requiring new infrastructure. The SDK replaces traditional workflow engines by using normal co…

Vercel redesigns deployment filters with natural language and type-ahead

Vercel updated its Deployments page with redesigned filters that support suggested options, type-ahead search, and natural language queries to speed up deployment discovery. The changes are live now f…

Vercel adds Cursor support to AI SDK harness layer via new adapter

Vercel introduced an official `@ai-sdk/harness-cursor` adapter, enabling Cursor to integrate with its AI SDK harness layer. This allows developers to switch between coding agents like Cursor, Claude C…

Pricing updated for Vercel: - Hobby: now includes 100 GB bandwidth/month, 10 GB storage/month, 10K requests/month - Pro: promotion changed — Includes $20 monthly credit to use across resources - Pro:

Pricing updated for Vercel: - Pro: allowance changed from $20 monthly credit to Included credit: $20/month

Vercel adds GLM 5.3 Flash to AI Gateway with multimodal support

Vercel now offers GLM 5.3 Flash from Z.ai on its AI Gateway, a multimodal model supporting text and vision input with a 1M token context window. It includes function calling, structured output, and st…

Vercel adds routing rules support for Python projects

Vercel now allows Python projects (FastAPI, Django, Flask) to define routing rules via its CDN, enabling rewrites and header modifications without redeploying. Rules can be managed via dashboard, CLI,…

Vercel adds Alibaba’s Qwen 3.8 Flash to AI Gateway

Vercel’s AI Gateway now supports Alibaba’s Qwen 3.8 Flash, a high-capacity model handling 1M-token context and 65k-token responses. It targets coding, tool use, and multi-step agent workflows, with di…

Vercel adds Meta's Muse Image to AI Gateway

Vercel integrated Meta's Muse Image model into its AI Gateway, enabling image generation and editing via a unified API. Users can generate images from text prompts or edit existing images with instruc…

Vercel Security Dashboard now generally available across all plans

Vercel launched its Security Dashboard as generally available, consolidating security posture visibility across accounts and projects. The tool flags misconfigurations like missing 2FA, long-lived cre…

Vercel adds Google's Gemini 3.5 Transcribe to AI Gateway

Vercel integrated Google's Gemini 3.5 Transcribe into its AI Gateway, offering both batch and live transcription via WebSocket. The model supports 85+ languages, dynamic language switching, and custom…

Vercel AI Gateway expands to 329 models with multimodal and video capabilities

Vercel’s AI Gateway now supports 329 models across chat, embedding, image, and video modalities, including recent additions like Alibaba’s Qwen3 series, Anthropic’s Claude 4/5 family, and video models…

Pricing updated for Vercel: - Hobby: allowance changed from 1 project, 10 GB bandwidth/month, 100 GB-hours/month to Limited usage caps - Pro: allowance changed from 12 environments, 100 GB bandwidth/m

Vercel Chat SDK adds Slack Enterprise Grid support

Vercel's Chat SDK Slack adapter now supports Slack Enterprise Grid, enabling org-wide bot installations with correct token resolution, tenant-scoped caching, and event deduplication. Enterprise ID fie…

Vercel offers free MiniMax M3 and M2.7 on AI Gateway until September 6, 2026

Vercel made MiniMax M3 and M2.7 models free to use on AI Gateway via GMI Cloud until September 6, 2026, using special -free model IDs. After the period, users must switch to paid access or configure f…

Vercel adds Wan 3.0 video model to AI Gateway

Vercel launched Wan 3.0 from Alibaba on its AI Gateway as a unified video generation model covering text-to-video, image-to-video, and reference-based generation. The model supports up to 30-second cl…

Vercel AI Gateway adds asynchronous video generation with webhooks and polling

Vercel’s AI Gateway now supports asynchronous video generation to avoid request timeouts during long-running video tasks. Developers can use webhooks, polling, or delayed retrieval via new SDK methods…

Vercel Connect reaches GA with scoped tokens and 100+ preset connectors

Vercel launched Vercel Connect in GA, replacing long-lived secrets with short-lived, scoped tokens issued at runtime via Vercel OIDC identity. The service ships with 100+ preset connectors for tools l…

Vercel protects Next.js apps from August 2026 security vulnerabilities

Vercel mitigated two critical Next.js vulnerabilities disclosed in August 2026: an AVIF-based remote code execution flaw and a Windows filesystem vulnerability. Vercel-hosted apps are protected withou…

Vercel Connect GA: secure, scoped tokens for agents replace long-lived credentials

Vercel launched Vercel Connect generally available, replacing long-lived tokens with short-lived, scoped credentials requested at runtime for agent-based applications. The system supports 100+ connect…

Vercel launches Run SDK for secure agent code execution

Vercel introduced the Run SDK, a package for executing untrusted JavaScript and TypeScript in a secure sandbox without direct access to application secrets or services. The SDK enables agents to write…

Vercel adds free tier to Speed Insights with expanded limits

Vercel introduced a free tier for Speed Insights, providing real-user performance data with 10,000 events per team every 30 days across all plans. The paid tier, now called Speed Insights Plus, retain…

Vercel adds Notion adapter to Chat SDK for agent collaboration

Vercel introduced a Notion adapter for its Chat SDK, allowing AI agents to participate in Notion comment discussions without separate codebases. The adapter maps Notion pages to channels and comment t…

Vercel Chat SDK adds XChat adapter for encrypted bot conversations

Vercel’s Chat SDK now includes an official XChat adapter enabling developers to build bots for end-to-end encrypted 1:1 and group chats on XChat. The adapter automates encryption, key management, and …

Vercel Connect adds managed Linq connector for iMessage, RCS, and SMS

Vercel introduced a managed Linq connector in Vercel Connect, enabling apps and agents to send and receive messages via iMessage, RCS, and SMS without managing credentials. The connector integrates wi…

Pricing updated for Vercel: - Hobby: allowance changed from 1 project, 10 GB Fast Data Transfer/month, 100 GB-hours of Serverless Function execution/month, 100 GB-hours of Edge Function execution/mont

Vercel Sandbox gains global region support with failover capabilities

Vercel Sandbox now supports global region deployment across four initial regions (Washington, D.C., San Francisco, Cleveland, Paris) to reduce latency for database and storage access. Pro and Enterpri…

Vercel elastic build machines now use Turborepo cache hits to prevent downgrades

Vercel updated its elastic build machines to consider Turborepo cache hits when determining machine size, preventing downgrades for warm-cache builds. This avoids potential failures in subsequent cold…

Vercel AI Gateway adds OIDC and BYOK support for authentication

Vercel’s AI Gateway now supports OpenID Connect (OIDC) tokens for automatic authentication on Vercel deployments alongside existing API keys. The Bring Your Own Key (BYOK) feature allows teams to use …

Vercel updates environment variables with Config and Secret types

Vercel replaced the Sensitive toggle with Config and Secret types for environment variables. Config values remain readable, while Secrets are hidden after saving. Existing Sensitive variables automati…

Vercel enables 5GB large functions and 30-minute durations for Bun runtime

Vercel expanded Bun runtime support on Vercel Functions to include large functions (up to 5GB) and extended max duration (up to 30 minutes) for Pro and Enterprise teams. Both features require Fluid co…

Pricing updated for Vercel: - Hobby: now includes 1 project, 10 GB Fast Data Transfer/month, 100 GB-hours of Serverless Function execution/month, 100 GB-hours of Edge Function execution/month - Pro: p

Vercel launches Deployment Storage for instant rollbacks and retention control

Vercel introduced Deployment Storage to keep deployment files available for inspection and instant rollbacks, enabling teams to revert to previous versions in seconds without rebuilds. The feature inc…

Pricing updated for Vercel: - Pro: allowance changed from 100 GB bandwidth/month, 10 GB storage/month, 1M ISR reads/month, 10K Simple Operations/month, 12 environments, 1 project, 1 concurrent sandbox

AI Social Club: Stockholm

Luma announced a private AI Social Club event in Stockholm featuring Vercel leadership discussing agent architecture and agentic enterprise trends. The closed-door gathering targets senior AI executiv…

AI Social Club: London

Vercel is hosting a private AI Social Club event in London for senior executives to discuss agent architecture and AI product shipping. The agenda includes a talk by Vercel leadership and a fireside c…

AI Social Club: Madrid

Luma announced a private AI Social Club event in Madrid featuring Vercel leadership discussing agent architecture and AI product shipping strategies. The closed-door forum targets senior AI executives…

AI Social Club: Paris

Vercel is organizing a private AI Social Club event in Paris for senior AI executives, focusing on agent architecture and infrastructure. The agenda includes a talk by Vercel leadership and a fireside…

Vercel CLI expands DNS, domain, and project command support

Vercel’s CLI now includes dedicated commands for managing DNS records, domain renewals, project state, observability, and team memberships. These features mirror dashboard/API capabilities and support…

Ora integrates Vercel’s eve framework and platform for AI agent benchmarking

Ora, an AI agent benchmarking platform, runs its entire testing infrastructure on Vercel, including the agent runtime and front-end/back-end components. The company uses Vercel’s eve framework as one …

Vercel introduces Claim Deployments for secure project ownership transfers

Vercel launched Claim Deployments, a feature enabling users to take ownership of deployments by generating a claim URL for others to assume control. This supports AI-generated deployments and simplifi…

Vercel introduces always-on tracing for production and preview traffic

Vercel launched always-on tracing in beta, enabling continuous collection of traces from live production and preview traffic without manual reproduction. Users control sampling via rules scoped to env…

Vercel adds DeepSeek V4 Flash Vision Experimental to AI Gateway

Vercel launched DeepSeek V4 Flash Vision Experimental on its AI Gateway, enabling multimodal inputs (text + images) for tasks like image description, OCR, and chart analysis. The model supports JPEG, …

Vercel v0 apps gain 100+ third-party integrations via Vercel Connect

Vercel introduced Vercel Connect, enabling v0-built apps to securely connect to over 100 third-party services like Slack, Google, Notion, GitHub, and Salesforce. Connectors are reusable across apps, m…

Vercel cuts GPT-5.6 Sol pricing 50% via AI Gateway through September 18

Vercel reduced list pricing for OpenAI’s GPT-5.6 Sol model by 50% on its AI Gateway, with input token rates dropping 20% and output rates falling a third. The discount applies to all service tiers and…

Pricing updated for Vercel: - Hobby: allowance changed from 1 project, 10 GB storage/month, 1 GB bandwidth/month, 10K Edge Function invocations/month, 1K ISR reads/month, 1 environment to 100 GB bandw

Vercel Functions now supports Bun 1.4 with Rust rewrite and WebSocket support

Vercel Functions added support for Bun 1.4, featuring a Rust rewrite, 2,900+ issue fixes, and 1,500+ additional Node.js compatibility tests. The update enables WebSocket support, Active CPU pricing, a…

Vercel CLI adds vcr command group to manage Container Registry

Vercel introduced a new `vercel vcr` CLI command group enabling direct management of Vercel Container Registry, including authentication, image builds, and repository operations. Users can now push im…

Vercel adds custom metrics support in Observability

Vercel now allows users to emit custom metrics from Vercel Functions using the `metric()` function, integrating them with built-in observability data. Metrics can be queried via a query builder, noteb…

Vercel Agent now available in Slack code channels

Vercel launched Vercel Agent in Slack code channels, enabling teams to collaborate with the coding agent directly in Slack. Users can create code channels, follow Agent’s work, review code, and approv…

Vercel CLI adds comment management for Vercel Toolbar

Vercel introduced a new CLI command, `vercel comments`, enabling full triage of Vercel Toolbar comments directly from the terminal. Users can list, inspect, reply, resolve, edit, or delete comments, w…

Vercel v0 introduces secure Snowflake authentication proxy to prevent OAuth token exposure

Vercel’s v0 Snowflake integration now uses a server-side proxy to authenticate Snowflake requests from AI-generated code without exposing user OAuth tokens. The proxy rewrites requests outside the san…

Pricing updated for Vercel: - Hobby: allowance changed from Limited usage caps to 1 project, 10 GB storage/month, 1 GB bandwidth/month, 10K Edge Function invocations/month, 1K ISR reads/month, 1 envir

Vercel adds Fish Audio models to AI Gateway with 30-day free offer

Vercel integrated Fish Audio’s text-to-speech and speech-to-text models into its AI Gateway, offering them for free through September 18, 2026. Four models are available, including low-latency streami…

Vercel integrates Agent with Slack for AI-powered infrastructure management

Vercel launched Vercel for Slack, integrating its AI agent into Slack channels to diagnose issues, review PRs, and execute approved changes directly within conversations. Available in Public Beta for …

Vercel Chat SDK adds Instagram adapter for bot development

Vercel introduced an Instagram adapter for its Chat SDK, enabling bots to send/receive DMs, media, and handle reactions via Meta's Instagram Messaging API. The adapter requires a Business or Creator a…

Vercel Connect adds managed Microsoft connector for Teams, OneDrive, and Graph API

Vercel introduced a managed connector for Microsoft in Vercel Connect, enabling apps and agents to access Microsoft products like Teams and OneDrive without manual app registration. The connector hand…

Vercel adds Algolia to its Marketplace for hosted search

Vercel integrated Algolia into its Marketplace, allowing users to provision and manage Algolia directly from Vercel. The integration supports indexing product catalogs, documentation, and websites wit…

Vercel for Slack public beta launches with AI Agent integration

Vercel launched Vercel for Slack in public beta, integrating its AI Agent into Slack channels. Users can now @mention @Vercel to investigate incidents, review code, or take approved actions like rollb…

Vercel KMS enables JWT signing from Functions without managing private keys

Vercel launched Vercel KMS, a managed asymmetric signing service for JWTs and raw messages from Vercel Functions, eliminating the need to handle private keys in code. Projects can create, rotate, and …

Vercel launches $1M bug bounty to test Sandbox isolation security

Vercel introduced a two-week public HackerOne program offering up to $1M in bounties to researchers who can escape its Vercel Sandbox isolation. The challenge targets vulnerabilities in both compute a…

Vercel for Platforms adds GitHub repository deployment support

Vercel for Platforms now allows teams to create deployments directly from users' GitHub repositories by passing a temporary, read-only GitHub access token with the gitSource parameter. This eliminates…

Vercel adds Cline harness adapter to AI SDK

Vercel introduced an official harness adapter for Cline, enabling it to run through the unified AI SDK harness layer. The adapter, `@ai-sdk/harness-cline`, allows developers to switch coding-agent run…

Vercel adds GLM 5.3 to AI Gateway with improved software engineering and security capabilities

Vercel’s AI Gateway now supports GLM 5.3 from Z.ai, offering better performance in complex software engineering and multi-step agent tasks while reducing output tokens. The model excels in vulnerabili…

Vercel integrates Cursor Origin repositories with Preview deployments in public beta

Vercel now supports connecting Cursor Origin repositories to Vercel for Pro customers, enabling automatic Preview deployments on pull requests and Production deployments on merges. The integration is …

Vercel slashes GPT-5.6 Sol pricing 50% via AI Gateway through September 2026

Vercel is offering a 50% discount on OpenAI’s GPT-5.6 Sol model through its AI Gateway until September 18, 2026, across all service tiers and regions. The discount applies automatically to requests ro…

Vercel CDN adds Encrypted Client Hello (ECH) support

Vercel CDN now supports Encrypted Client Hello (ECH), encrypting the Server Name Indication (SNI) in TLS handshakes for domains using Vercel DNS. This prevents network observers from seeing which host…

Vercel adds Grok Build harness adapter to AI SDK layer

Vercel introduced a new harness adapter for Grok Build, enabling developers to run Grok Build coding agents through the unified AI SDK harness layer. The adapter, `@ai-sdk/harness-grok-build`, integra…

Vercel offers free GLM 5.2 for eve agents through August 27

Vercel is providing free access to Z.ai's GLM 5.2 (1M-token context) for eve agents until August 27, 2026, via Blackbox AI on Vercel's AI Gateway. New and existing eve agents can use GLM 5.2 as the de…

Vercel AI SDK harness layer now supports ACP-compatible harnesses

Vercel introduced `@ai-sdk/harness-acp`, a new package enabling the AI SDK harness layer to work with any Agent Client Protocol (ACP)-compatible harness via a meta adapter. Previously, harness adapter…

Vercel adds Google’s Gemini 3.7 Flash to AI Gateway at 50% off

Vercel launched Google’s Gemini 3.7 Flash on its AI Gateway platform with a 50% discount through December 31, 2026. The model improves software engineering and agentic workflows by reducing failed loo…

Vercel launches official MCP server for AI tool integration

Vercel introduced Vercel MCP, an official Model Context Protocol (MCP) server enabling AI assistants to securely interact with Vercel projects. The integration supports tools for searching docs, manag…

Vercel highlights Winter 2026 intern cohort shipping production features

Vercel's Winter 2026 intern cohort shipped production features across CDN, v0, financial infrastructure, and AI Gateway, gaining full ownership of projects used by millions. Interns contributed to ent…

Vercel adds one-click upgrade for deprecated Node.js versions

Vercel introduced a one-click upgrade feature in its dashboard to migrate projects from Node.js 20 or earlier to Node.js 24. The change updates the Node.js version in project settings, though projects…

Exa joins Vercel Agent Marketplace with neural search integration

Vercel added Exa’s neural search engine to its Agent Marketplace, enabling developers to integrate real-time, high-quality web search into AI agents with a single API key. The integration supports ins…

Vercel extends free domain offer to all Pro teams

Vercel expanded its free domain offer to all existing Pro teams, covering one eligible domain per team across six TLDs for the first year. The perk renews at standard pricing after the first year.

Vercel builds AI SDK software factory to automate 25-35% of PRs and 70-80% of issues

Vercel introduced an internal software factory for its AI SDK to address a backlog of over 1,000 open issues and 800 PRs. The factory autonomously processes issues and PRs, generating 25-35% of merged…

Vercel integrates Exa web search into AI Gateway and eve agents

Vercel made Exa web search free through August 31 for AI Gateway users and set it as the default search tool for eve agents. The integration eliminates the need for a separate Exa API key, routing req…

Vercel adds one-command coding agent setup via AI Gateway

Vercel introduced a new command, `vercel ai-gateway coding-agents setup`, to centralize coding agent configuration across 9 supported agents (e.g., Claude Code, Codex, OpenCode) using AI Gateway. The …

Vercel AI Gateway updates DeepSeek V4 Pro weights with new model ID

Vercel's AI Gateway now runs DeepSeek V4 Pro with updated weights by default, accessible via the new model ID `deepseek/deepseek-v4-pro-0813`. Developers must update their code to use the new ID, whil…

Vercel adds Grok 4.6 to AI Gateway with 500K token context window

Vercel’s AI Gateway now supports Grok 4.6 from SpaceXAI, featuring a 500K token context window, multi-modal (text/image) inputs, and configurable reasoning levels. The model is accessible via the AI S…

Vercel migrates build warm pool state from Redis to DynamoDB

Vercel migrated the state tracking its build warm pool—used for every deployment—from Redis to DynamoDB to ensure durability, as billing mappings could not be rebuilt if lost. The live migration ran i…

Vercel CLI now supports 100+ service connectors via Vercel Connect

Vercel introduced CLI-based setup for 100+ third-party service connectors through Vercel Connect, replacing partial terminal/dashboard workflows. Users can now create and attach connectors directly fr…

Vercel Connect adds observability support for token lifecycle tracking

Vercel Connect now provides line-level visibility into token lifecycles, including creation, usage, and status, with runtime events, correlation IDs, and activity filters. Observability is available o…

Vercel Marketplace adds LaunchDarkly for feature flag management

Vercel integrated LaunchDarkly into its Marketplace, enabling developers to deploy feature flags directly through Vercel's Global Config and Flags SDK. Users can sync flags, target releases, run exper…

Vercel Enterprise Managed Users now generally available

Vercel launched Enterprise Managed Users (EMU) as generally available, giving organizations full control over accounts tied to verified domains via SAML SSO and SCIM provisioning. Managed accounts are…

Vercel warns of AI-driven cybersecurity threats and launches deepsec tool

Vercel highlights that open-weight AI models like Kimi K3 can perform offensive security research today, while frontier models (except Fable 5) can aid defensive work. They introduce deepsec, an open-…

Vercel introduces network boundary controls for Sandbox to prevent data exfiltration

Vercel announced that its Sandbox now enforces network boundary controls alongside compute isolation to prevent untrusted code from exfiltrating data or probing internal services. The new firewall pol…

Pricing updated for Vercel: - Hobby: now includes Limited usage caps - Pro: now includes Pay-as-you-go, uncapped usage beyond included credit - Enterprise: promotion changed — Contact sales team - Ent

Vercel simplifies deepsec onboarding with single-command setup

Vercel’s open-source security review tool, deepsec, now supports one-command onboarding via `npx deepsec init`, automating repository setup, dependency installation, model access configuration, codeba…

Vercel Sandbox migrates to Vercel Managed Images with Ubuntu base

Vercel introduced Vercel Managed Images (VMI) to replace deprecated Sandbox runtimes, defaulting to the new `vercel/sandbox/universal:latest` image on Ubuntu 26.04 with Node.js 24, Python 3.14, and co…

Vercel Functions Bun runtime now supports Bun.serve as entrypoint

Vercel added support for `Bun.serve()` as a function entrypoint in its Bun runtime, enabling direct deployment of routes-based servers and WebSocket handlers without framework wrapping. Developers can…

Vercel adds Grok Imagine Image 2.0 Preview to AI Gateway

Vercel now offers Grok Imagine Image 2.0 Preview from xAI on its AI Gateway, enabling high-fidelity image generation with improved typography, layout planning, and image editing. Developers can access…

Pricing updated for Vercel: - Pro: new promotion — $20 included credit to use across resources - Pro: added annual billing at $16/mo (20% off) - Enterprise: new promotion — Contact our sales team to l

Pricing updated for Vercel: - Pro: promotion ended (was: $20 included credit to use across resources) - Pro: allowance changed from 1 user to 1 developer seat included - Enterprise: promotion ended (w

Vercel AI Gateway and Vercel Sandbox now integrate with Hermes Agent

Vercel AI Gateway is now available as an inference layer for Hermes Agent, enabling access to 200+ models at no markup on tokens. Hermes Agent can also run commands in an isolated Vercel Sandbox micro…

Vercel Container Registry adds public repository visibility

Vercel Container Registry now supports making repositories public, granting read-only access to all Vercel teams instead of a limited set. Public images can be pulled by any team and used in Vercel Sa…

Vercel adds free domain for first year with new Pro subscriptions

Vercel now includes a free domain for the first year with new Pro subscriptions, covering six TLDs (.online, .site, .space, .store, .tech, .website). Teams have 30 days to claim one domain after subsc…

Vercel Audit Log Drains add Datadog, Splunk, and Panther support

Vercel expanded Audit Log Drains to support Datadog, Splunk, and Panther as destinations for enterprise customers, replacing the previous Custom SIEM Log Streaming feature. Users can now forward audit…

Vercel Sandbox integration with Herdr enables isolated agent terminals

Vercel launched a plugin enabling terminal coding agents like Claude Code, Codex, and OpenCode to run in isolated Vercel Sandboxes via Herdr, a tmux-style manager. Agents operate in remote panes witho…

Vercel adds Ling 3.0 Tiny to AI Gateway as free model until August 13

Vercel’s AI Gateway now hosts Ling 3.0 Tiny, a 7.9B-parameter MOE model with 256K context and 32K output tokens, replacing Ling 3.0 Flash in the free slot. It supports function calling, prompt caching…

Vercel adds durable approvals to Chat SDK for workflow suspensions

Vercel introduced a new `chat/workflow` subpath in its Chat SDK enabling human-in-the-loop approvals via `requestApproval`, which suspends Workflow SDK processes until a decision is made. The feature …

Vercel launches Agent Plugins 1.0.0 as open standard for AI agent extensions

Vercel introduced Agent Plugins 1.0.0, an open, vendor-neutral standard for packaging AI agent Skills and MCP servers into portable plugins. The format simplifies distribution and discovery across cli…

Vercel AI Gateway adds Seedance 2.5 video model from ByteDance

Vercel AI Gateway now supports Seedance 2.5, a ByteDance video generation model that creates up to 30-second clips with seamless continuity, supporting text, video, image, and audio inputs. It enables…

Vercel CLI now auto-installs agent skills for Marketplace integrations

Vercel CLI now automatically installs agent skills from skills.sh when users add a Vercel Marketplace integration, enabling agents to use the integration immediately. Users can discover integrations v…

Vercel launches Agent Plugins 1.0.0 as open standard for agent skills

Vercel introduced Agent Plugins 1.0.0, an open, vendor-neutral standard for packaging Agent Skills and MCP servers into portable plugins. The format includes a root manifest and fixed component locati…

Vercel Sandbox raises concurrent sandboxes to 10,000 and vCPU rate to 5,000/min

Vercel increased default quotas for Vercel Sandbox on Pro and Enterprise plans, raising concurrent sandboxes from 2,000 to 10,000 and vCPU allocation rates up to 5,000 per minute via a new dynamic quo…

Vercel AI Gateway now exports OpenTelemetry traces via Drains

Vercel’s AI Gateway now generates OpenTelemetry traces for every request, which Pro and Enterprise teams can export to OTLP/HTTP-compatible endpoints using Vercel Drains. Traces include model routing,…

Vercel launches v0 API for programmatic app-building agent access

Vercel introduced a new v0 API enabling programmatic, headless access to its v0 app-building agent, allowing users to generate, edit, and deploy apps via API calls. The API supports real-time streamin…

Vercel introduces new domain setup page with live registration tracking and deployment paths

Vercel added a post-domain-checkout setup page that tracks domain registration in real time and offers direct paths to deploy new projects, connect existing ones, proxy/redirect existing sites, or set…

Vercel AI Gateway now available on AWS Marketplace

Vercel announced the availability of its AI Gateway on AWS Marketplace, allowing teams to purchase and consolidate inference spend through their AWS bill. The offering includes private offers with ann…

Vercel adds custom project avatars in dashboard and GitHub bot comments

Vercel now lets users set custom project avatars in the dashboard, replacing the automatic favicon-based avatars. The feature applies to all projects, including microfrontends with shared favicons and…

Vercel adds step timing measurement in Workflows trace viewer

Vercel introduced a feature to measure time between steps in Workflows using the trace viewer. Users can now select a step and hover over another while holding a modifier key to see a measurement line…

Vercel Sandbox firewall features now available on Hobby plan

Vercel extended its full Sandbox firewall features to the Hobby plan, enabling free-tier users to control outbound network traffic with domain/IP-based rules. Secrets can be attached to requests witho…

Vercel AI Gateway adds Meta’s Muse Spark 1.2 coding model

Vercel AI Gateway now supports Meta’s Muse Spark 1.2, a coding-focused model with improved code generation, debugging, and repository-scale workflows. Users can access it via the AI SDK or coding agen…

Vercel adds agent-browser extension to eve platform

Vercel’s eve agent now supports browser automation via the `@agent-browser/eve` extension, enabling agents to navigate, interact with, and inspect web pages within a sandboxed Chromium environment. Th…

Vercel Sandbox integrates with Devin Outposts for agent-based workflows

Vercel Sandbox now supports Devin Outposts, enabling isolated Devin sessions to run in Vercel’s Sandbox microVMs with state preservation via snapshots and network policy enforcement through firewall r…

Vercel deploys up to 33% faster for apps with many ISR pages

Vercel improved deployment speed by up to 33% for apps using Incremental Static Regeneration (ISR) by bundling routing metadata with prerendered pages instead of splitting it into a separate upload. T…

Vercel fully supports Next.js 16.3 with immutable static assets and faster routing

Vercel announced full support for Next.js 16.3, enabling immutable static assets by default and delivering significant performance gains. Upgraded apps saw 45% fewer prefetch requests, 17% fewer CDN r…

Vercel launches skill packs for agent skills on skills.sh

Vercel introduced skill packs on skills.sh, allowing users to bundle multiple agent skills into shareable packs accessible via URL or GitHub organization. Packs can be created from community skills or…

Factory scales backend to 1B daily requests on Vercel’s Next.js platform

Factory, an enterprise software platform, scaled its cloud backend to one billion daily API requests using a single Next.js backend on Vercel. The setup handles API routes, middleware, and webhooks wi…

Vercel WAF for Blob reaches general availability

Vercel announced the general availability of Vercel WAF for Blob, a firewall protection layer for Blob storage that blocks unwanted traffic at the edge using custom rules. The feature is now supported…

Vercel Container Registry adds team repository sharing

Vercel Container Registry now supports sharing repositories across teams with read-only access, enabling up to 100 teams per repository. Shared images can be used in Vercel Sandbox via team-scoped ref…

Factory scales backend to tens of millions of daily requests on Vercel

Factory, an enterprise software platform, scaled its Next.js-based backend on Vercel to handle tens of millions of daily API requests with 350ms p95 response times. The company leveraged Vercel’s serv…

Vercel adds Qwen 3.8 Max to AI Gateway

Vercel AI Gateway now supports Alibaba’s Qwen 3.8 Max, a 2.4-trillion-parameter model handling text-only and vision-language tasks with a 1M-token context window. The model targets software engineerin…

Pricing updated for Vercel: - Pro: allowance changed from 100 GB bandwidth/month, 10 GB storage/month, 1K ISR reads/month, 1 project, 12 environments, 10 concurrent sandboxes, 5K events/month, 1M API