Vercel — Cloud, Developer Tools product updates and releases, tracked on megachangelog.
Vercel shares how its intern cohort contributed to major products over four months, including CDN routing rules, v0 features, billing infrastructure, and AI systems. The interns shipped features used by millions and gained real ownership and autonomy working on production systems.
Team owners and members can now upgrade all projects using Node.js 20 or earlier to Node.js 24 directly from the Vercel Dashboard with a single click. The upgrade updates project settings but requires manual updates if a version is also defined in package.json.
GLM 5.2, an open-weights coding model with 1M-token context, is now free for eve agents through August 27 when served by Blackbox on AI Gateway. After the promotion ends, the model remains available at standard provider rates.
Google's Gemini 3.7 Flash model is now available on Vercel's AI Gateway with a 50% discount through December 31st, 2026. The model improves on software engineering and agentic tasks with better reliability and generates desktop and web application code from design mocks.
Grok Build is now integrated into the AI SDK harness layer, allowing developers to run Grok Build through a unified interface alongside other supported coding-agent runtimes without changing application code.
The AI SDK harness layer now supports any Agent Client Protocol (ACP)-compatible harness through the new @ai-sdk/harness-acp package, enabling integration with harnesses that don't have direct adapters. This meta-adapter wraps the ACP protocol instead of specific runtimes, letting you build an adapter for any harness that ships an ACP-compatible package.
Exa's neural search engine is now available as a native integration in the Vercel Agent Marketplace, enabling developers to add intelligent search and research capabilities to their apps with a single API key and unified billing through Vercel.
DeepSeek V4 Pro now runs on updated weights (dated 0813) through AI Gateway. Existing requests automatically use the new weights without code changes, and you can explicitly target the updated model by specifying the deepseek/deepseek-v4-pro-0813 identifier.
Exa web search is now available for free through August 31 on AI Gateway and Eve, with no separate API key required. The tool is now the default web search for Eve agents and works with any AI Gateway model, returning web results with domain and date filtering support.
Grok 4.6 from SpaceX AI is now available on Vercel's AI Gateway with a 500K token context window, support for text and image inputs, and configurable reasoning levels. Use it by setting the model to xai/grok-4.6 in the AI SDK or coding agents.
Vercel AI Gateway now supports routing for 9 different coding agents including Claude Code, Codex, and OpenCode. Developers can set up all agents in one command with centralized observability, billing, and policy controls across 200+ models.
Vercel built an autonomous software factory for the AI SDK open-source project that uses specialized agents to classify issues, reproduce bugs, implement fixes, and manage pull requests. The factory currently authors 25-35% of merged PRs and closes 70-80% of issues while keeping humans in control of all final merges, using isolated sandboxes for security and evidence-based assessments to guide reviewer efficiency.
All Vercel Pro teams can now claim a free domain for one year across six TLDs. After the first year, domains renew automatically at regular pricing.
Vercel migrated the state management for the build warm pool from Redis to DynamoDB to provide durable storage for billing mappings while maintaining production availability. The migration used phased rollouts with feature flags and comprehensive validation to safely transition traffic without downtime.
Vercel Connect now provides line-level visibility into the token lifecycle, showing who created tokens, which app or project used them, and their current status. The new Observability tab includes runtime events, correlation IDs for system matching, and activity filters, with retention periods based on plan level and options to forward events to custom webhooks on Pro and Enterprise plans.
Enterprise Managed Users (EMU) is now generally available, giving organizations full control over Vercel accounts on verified domains through their identity provider. Managed users authenticate via SAML SSO only, with automatic account provisioning and lifecycle management through Directory Sync.
Vercel Connect now allows you to set up and integrate 100+ services entirely through the CLI. Previously, you could only complete terminal setup for some services and had to use the dashboard for others. The CLI automatically pre-populates connector details like brand name and auth type, then guides you through credential setup.
Vercel shares insights on AI-enabled defensive cybersecurity, highlighting how frontier models like OpenAI's Sol 5.6 can be used today for vulnerability discovery and defense, and introduces deepsec, an open-source security harness for scanning large codebases before open-weight models catch up in capabilities.
Vercel's August 2026 AI Gateway Production Index shows DeepSeek surpassed Google as the second-largest AI model by token volume, while average cost per token fell 13.6% due to widespread adoption of cheaper open-weight and frontier models. Open-weight models now represent over a third of all tokens on the gateway at significantly lower rates.
Vercel Sandbox now includes a host-based firewall that controls network egress and injects credentials at the boundary, preventing untrusted code from exfiltrating data, scanning internal networks, or misusing stored credentials. The firewall supports domain allowlists, CIDR policies, temporary connectivity windows, and transparent TLS inspection without decryption for authorized connections.
LaunchDarkly feature flags integration is now available on the Vercel Marketplace, enabling developers to set up feature flags quickly without additional configuration. Users can sync flags into Global Config, run experiments with rollback capabilities, and manage flags via the Vercel Toolbar.
Vercel Sandbox now uses Vercel Managed Images (VMI), versioned open-source base images that replace deprecated runtimes. The new universal image defaults to Ubuntu with Node.js, Python, coding agents, and standard utilities preinstalled, with nightly releases and automatic security updates.
The deepsec open-source security review tool now offers one-command setup and improved initialization that automatically creates the isolated workspace, configures model access, generates codebase analysis, and runs initial scans. Setup is checkpointed at each step so interrupted runs can resume from where they left off.
Vercel Functions now support Bun.serve as a native entrypoint, allowing you to deploy servers with WebSocket support as-is without framework wrapping. Servers use Fluid compute with per-message pricing for WebSocket connections.
Grok Imagine Image 2.0 Preview from xAI is now available on Vercel's AI Gateway. The model excels at detailed instruction following, maintaining structure in complex layouts like infographics and posters, and supports image editing with consistent subject generation across repeated calls.
Hermes Agent can now use Vercel AI Gateway as its inference layer with access to 200+ models at no token markup, and run agent commands in isolated Vercel Sandbox microVMs. Both features integrate into the setup wizard with live model availability and pricing information, and Sandbox supports Node 24, Node 22, and Python 3.13 runtimes.
Vercel Container Registry now supports making repositories public, allowing anyone with a Vercel account to pull and use images while maintaining read-only access. Repository visibility can be managed from the project dashboard, CLI, or REST API.
Audit Log Drains now stream team audit events to Datadog, Splunk, and Panther in addition to custom HTTPS and Amazon S3 endpoints. This replaces the Custom SIEM Log Streaming feature and is available on Enterprise plans.
New Pro subscriptions now include one free domain for the first year across six popular TLDs, with 30 days to claim it at checkout or from the dashboard. Domains renew automatically at regular pricing after the first year.
Terminal coding agents like Claude Code, Codex, and OpenCode can now each run in isolated Vercel Sandboxes, orchestrated through a tmux-style manager that displays multiple agents in side-by-side panes. Projects are uploaded to the Sandbox and agents run there without affecting your local machine, enabling safe parallel execution of multiple agents.
Chat SDK now supports human-in-the-loop approvals through a new requestApproval method that posts a card with Approve and Deny buttons and suspends workflow execution until a decision is made. The wait can last any duration and survives deploys and restarts, with support for scoped approvers, verified decisions, and an audit trail.
Seedance 2.5, a ByteDance video generation model, is now available on AI Gateway. It can generate up to 30 seconds of video with continuous camera movement and supports text, video, image, and audio inputs in a single request for more flexible creative workflows.
Agent Plugins 1.0.0 is now available as an open, vendor-neutral standard for packaging Agent Skills and MCP servers into portable plugins that compatible agent clients can discover and load. The standard defines a common manifest format and fixed component locations, giving each client full control over installation, distribution, and user experience.
When installing integrations from the Vercel CLI, provider agent skills are now automatically installed alongside the integration, enabling agents to use the integrated tools without manual configuration. Users can also discover and browse available integrations directly from the terminal.
Agent Plugins 1.0.0 is now publicly available as an open, vendor-neutral standard for extending AI agents with reusable skills and MCP servers. The standard defines a portable contract using a minimal plugin.json manifest, allowing plugin authors to package components once and work across multiple supporting clients including ChatGPT, Cursor, GitHub Copilot, and VS Code.
Ling 3.0 Tiny, a lightweight MOE model with 1.3B active parameters, is now available on Vercel's AI Gateway for free use through August 14. The model supports 256K token context windows, 32K output tokens, and native function calling, making it ideal for responsive agents and multi-turn conversations.
Users can now manually set custom avatars for their projects in the Vercel dashboard, replacing the automatic favicon detection. Avatars display throughout the dashboard and on Vercel bot comments in GitHub, enabling distinct branding for microfrontends and backend projects without favicons.
After purchasing a domain on Vercel, users are taken to a dedicated setup page that tracks domain registration in real-time and provides direct actions to deploy a new project, connect an existing one, proxy or redirect an external site, or configure email with DNS presets. The page displays registration progress and unlocks setup actions as the domain becomes ready.
AI Gateway now produces OpenTelemetry traces for every request and Pro/Enterprise teams can export them via Vercel Drains to any OTLP/HTTP-compatible endpoint, with native integrations for Braintrust, Dash0, Kubiks, Sentry, and Statsig. Traces include full request lifecycle details like model routing, fallback attempts, token usage, latency, and attribution metadata, with sampling controls and per-trace pricing of $0.05 per 1,000 traces.
Vercel has launched a new v0 API providing programmatic, headless access to v0's app-building agent. Users can send prompts to generate full apps with dev servers and preview URLs, with support for streaming, async operations, and integration into custom products and agent workflows.
AI Gateway can now be procured through AWS Marketplace, allowing teams to consolidate inference spend on their AWS bill with private offers and usage-based pricing. The service provides a single API to hundreds of models with built-in reliability, cost controls, and governance features.
Vercel has increased default quotas for 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 with a new dynamic quota system that scales with usage.
You can now search for and purchase domains directly from the Vercel dashboard with automatic nameserver configuration, eliminating the need to manually set DNS records.
Vercel Sandbox's egress firewall is now available on the free Hobby plan, enabling network isolation and credential management for all users. The firewall supports allow-all, deny-all, and custom network policies with domain and IP-based rules, attachable secrets to outbound requests, and live updates via Sandbox CLI.
Meta's Muse Spark 1.2 model is now available on Vercel's AI Gateway with improvements in code generation, debugging, codebase understanding, and end-to-end developer workflows. The model excels at long-horizon coding tasks like generating whole repositories and sustaining iterative development cycles.
You can now measure the time elapsed between any two steps in the trace viewer using a modifier key to select and hover steps, revealing measurement lines that help identify delays and queue waits. This feature is available in Vercel Observability and the Workflow SDK.
Vercel Sandbox now integrates with Devin Outposts, allowing each Devin session to execute in its own isolated Sandbox microVM. Teams can preserve session state with Sandbox snapshots and apply network policies through Sandbox firewall rules.
Vercel now fully supports Next.js 16.3 with leaner prefetching and immutable static assets. Key improvements include 17% fewer CDN requests for static content, up to 30% faster deployment times, 45% fewer prefetch requests, and enhanced observability for ISR and Partial Prerendering.
You can now bundle multiple agent skills into shareable packs on skills.sh and distribute them via single URLs or with your GitHub organization. Create packs from community skills or your own sources, and install or update them with simple commands.
Deploys are now significantly faster for apps using Incremental Static Regeneration (ISR) and Partial Prerendering. Routing metadata for ISR pages now stays bundled during upload instead of splitting into separate uploads, with the optimization applying automatically to all deploys.
Eve agents can now navigate the web and interact with web pages using the new agent-browser extension, which provides tools for page navigation, content reading, clicking, form filling, screenshots, and network inspection—all running in the agent's sandbox with security controls.
DeepSeek V4 Flash is available at 90% off on AI Gateway when routed through Novita for Vercel Pro customers through August 11. After the promotional period, the model remains available at standard rates without markup.
Vercel WAF for Blob is now generally available for production use on all plans, allowing you to protect blob stores with custom rules that match on IP address, country, path, and more. Rules are evaluated at Vercel's edge to deny, challenge, or rate-limit requests before data transfer occurs, preventing unwanted traffic and unnecessary billing.
Factory, an AI-powered software factory platform, demonstrates how to scale a cloud backend to tens of millions of daily requests using Vercel's Next.js infrastructure. The company runs its entire backend on Next.js with API routes, middleware, and webhook handlers, achieving 350ms p95 response times and enabling non-technical teams to autonomously build and deploy internal tools using Vercel's APIs, while protecting against threats with Vercel's Observability and Web Application Firewall.
Factory demonstrates how to run an entire cloud backend on Vercel's Next.js platform, handling one billion daily API requests across routes, middleware, and webhooks with 350ms p95 response times. The company uses Droids for self-serve deployments and leverages Vercel's observability and Web Application Firewall for security without dedicated infrastructure.
Vercel Container Registry now supports sharing repositories with other teams in read-only mode, allowing recipients to pull and use images while preventing pushes, deletions, or re-sharing. Repository sharing can be managed via the project dashboard or Vercel CLI and supports up to 100 teams per repository.
Qwen 3.8 Max, a multimodal model with 2.4 trillion parameters and 1 million token context window, is now available through Vercel AI Gateway for text, vision, software engineering, and productivity tasks. The model can be accessed via unified API with tracking, optimization, and no platform fees on inference.
AI Gateway spend budgets can now be scoped to teams or projects in addition to API keys. Set dollar limits on any scope, and the gateway enforces them by metering spend and rejecting requests once a limit is reached, with options for email alerts and default budget inheritance.
DeepSeek V4 Flash has been updated with new weights on Vercel's AI Gateway, delivering significantly stronger agentic capabilities and improved performance scores. The updated model is automatically available to existing requests with no code changes required.
Vercel Observability now supports structured search for workflow runs, allowing you to filter by workflow, environment, deployment ID, region, and custom attributes. Search suggestions are smart and based on your project data, and filters appear as removable chips in a shareable URL.
Vercel Passport, which protects deployments with your own identity provider (Okta, Microsoft Entra ID, OIDC), is now generally available. The service authenticates visitors through your provider, forwards verified identity tokens to your deployment, and supports advanced features like custom claims, token forwarding, automation bypass, and custom environments.
Laguna S 2.1 from Poolside now has 10x more capacity on AI Gateway for both paid and free versions, enabling higher-volume agentic coding and long-running tasks. Users can set the model parameter to poolside/laguna-s-2.1 or poolside/laguna-s-2.1-free in the AI SDK or configure it in coding agents via AI Gateway.
Vercel MCP now supports the 2026-07-28 MCP specification with a stateless request model and updated authorization behavior. Clients built for the 2025 protocol continue to work, and newer clients automatically adopt the new spec through the same endpoint.
AI Gateway now includes a dedicated logs page that displays all requests sent through the gateway with cost, token counts, duration, and routing details. Users can filter by provider, model, or status, search by request ID, export data as CSV or JSON, and view detailed request information including token breakdown, fallback attempts, and regional restrictions.
The Microsoft Teams adapter for Chat SDK now enables bots to add and remove reactions on messages and send ephemeral messages that only specified users can see. The release also adds support for author email retrieval, custom token factories, native DM status, improved conversation routing, and fixed mention conversion.
Vercel's CDN will now pass through Server-Timing response headers to clients starting August 10, 2026, allowing backend metrics like database query time and cache hits to appear in browser network panels and Performance API. Users can preserve the previous behavior by adding a transform in vercel.json.
The Sandbox SDK now supports multiple Linux users and groups, allowing agents to run side-by-side with isolated file systems and private home directories. Groups can create shared workspaces when collaboration is needed, making multi-agent systems easier to build.
Hydrogen, Shopify's headless storefront framework, has been rebuilt in partnership with Vercel as open source and runtime-agnostic. Developers can now use any JavaScript framework like Next.js, Nuxt, or Svelte, with built-in best practices including type-safe API clients and cart state management, enabling retailers to ship storefronts faster.
mcp-handler version 2.0 now supports the latest Model Context Protocol specification and MCP TypeScript SDK v2, with stateless Streamable HTTP compatibility and native MCP protocol support. Node.js 20 or later is required, and the deprecated HTTP+SSE transport has been removed.
Deployments are now up to 7 seconds faster end-to-end through platform optimizations that removed fixed overhead and CLI improvements. The improvements are most noticeable on smaller builds where orchestration overhead represents a larger share of total build time.
GPT-5.6 Luna and Terra models are now cheaper with reduced token pricing, and GPT-5.6 Sol's fast mode is now 2.5x faster. AI Gateway passes upstream rate changes directly to users with no code changes required for existing requests.
You can now create Vercel Access Tokens limited to a specific project, allowing you to authenticate with the Vercel API while restricting access to only resources within that project. This improves security by ensuring jobs, tools, and workflows can only access their designated projects.
Enterprise customers can now apply their Flexible Commitment toward eligible Marketplace resources, eliminating separate billing workflows and streamlining procurement of third-party services. The feature rolls out with initial partners including Neon, Supabase, and Redis, with more providers coming.
Turborepo and Vercel Remote Cache now support OIDC token exchange from CI/CD workflows, allowing short-lived access tokens instead of long-lived Personal Access Tokens. OIDC tokens are more secure, scoped only to Remote Cache, and associated with teams rather than individuals.
Inkling Small from Thinking Machines is now integrated into AI Gateway, offering comparable performance to larger models at a quarter of the size with lower compute costs. The model supports audio and image reasoning, tool use, and controllable thinking effort to balance quality against latency and cost.
MiniMax H3, a text-to-video and image-to-video model, is now available on Vercel's AI Gateway. The model generates 2K video from text prompts, starting images, keyframe pairs, or reference material including images, video, or audio.
Vercel now supports ChatGPT as an authentication option, allowing users to sign in with their ChatGPT account on Vercel and v0, including via the Vercel plugin for ChatGPT. Team security requirements like 2FA and SSO still apply.
AI Gateway now offers a unified fast mode abstraction that allows you to request lower latency or higher throughput across multiple models by setting speed to 'fast'. This feature automatically falls back to standard speed when fast tier isn't available, and is now in beta.
You can now discover and install integrations for eve agents directly from the CLI. Integrations come from the official eve catalog and third-party sources, with support for custom registries, and can add tools, channels, or full extensions to your project.
Pro and Enterprise teams can now buy extra custom environment capacity directly from the dashboard, API, or CLI in packs of 5 for $50/month each, without needing to contact sales. Custom environments help teams model their release process with named stages, branch tracking, environment variables, and domains.
Edge Config has been renamed to Global Config to better reflect its globally replicated nature with ~1ms reads across regions. All existing projects continue to work, though users should upgrade the @vercel/global-config SDK and update environment variable references from EDGE_CONFIG to GLOBAL_CONFIG. Storage limits have increased to 1 MB on all plans.
Grok Voice Think Fast 2.0, a speech-to-speech voice model from xAI, is now available on AI Gateway with improved reasoning, transcription accuracy, and lower latency for tool calls. The model reasons in parallel with speech and handles real-world audio conditions including background noise and telephony compression.
Vercel Sandbox now supports forking via the Sandbox.fork() method, allowing you to create branches from a shared base sandbox that inherit configuration and environment variables. Forks take similar time and have the same limits as creating a new sandbox.
Vercel Connect now enables linking connectors to custom environments, allowing deployments to request provider tokens and receive forwarded webhooks. Custom environments appear in the dashboard alongside Production, Preview, and Development, and are available on Pro and Enterprise plans.
AI Gateway now supports regional inference, allowing you to pin requests to US or EU data centers via the inferenceRegion field. Every response reports which region served the request, ensuring data residency and compliance requirements are met consistently across all model providers.
eve agents on Slack can now keep replies in a thread without repeated mentions, cancel in-progress responses, or reset conversations entirely. New hooks allow agents to react to any Slack event the app subscribes to, with helpers for detecting mentions, checking active sessions, and listing thread participants.
This case study highlights how Sandstone, an AI-powered legal workflow platform, achieved 40x revenue growth in under six months by leveraging Vercel's deployment speed, preview environments, Flags, and AI SDK to ship complex features and multi-step agentic workflows rapidly. Vercel's infrastructure enabled the startup to respond to customer requests in minutes and build composable AI-driven legal automation for enterprise customers.
Vercel released DeepsecBench, a benchmark that evaluates how well different AI models find cybersecurity vulnerabilities in application code. The benchmark includes recall, precision, cost, and timing metrics to help organizations choose the right combination of models for their security scanning needs based on budget and codebase complexity.
AI Gateway now supports WebSocket mode for the OpenAI Responses API, allowing persistent connections to send incremental input instead of re-sending full context on each turn. This enables up to 40% faster end-to-end execution for agentic rollouts with multiple tool calls.
You can now run Claude Managed Agents with Chat SDK, which provides a server-side agent loop including model, tools, and session state management. Chat SDK offers a type-safe chat interface with adapters for Slack, WhatsApp, Discord, Teams, and other platforms, enabling you to deploy agents across multiple channels with minimal setup.
Kimi K3 and Kimi K3 Fast models from Moonshot AI are now available on AI Gateway through US-based providers including Baseten and Fireworks, with support for Zero Data Retention (ZDR). Users can route requests to US data centers for compliance and leverage automatic failover across multiple providers for improved uptime and throughput.
Nuxt released versions 4.5.1, 3.21.10, and DevTools 3.3.1 to address eight security vulnerabilities including a high-severity server-side remote code execution flaw. Vercel deployed WAF mitigations platform-wide before public disclosure, automatically protecting deployed applications, but users must still upgrade to patched versions for complete protection.
Vercel WAF now protects Vercel Blob stores with the same rules (deny, challenge, rate limit) that guard deployments, evaluated at the edge with no code changes required. Protection is available on all plans and configured via a single dashboard switch during the beta period.
Workflow steps on Pro and Enterprise plans can now run for up to 30 minutes instead of 800 seconds by opting in with an environment variable. This requires Fluid compute and a supported Node.js or Python runtime, with Hobby plans remaining limited to 5 minutes.
Claude Opus 5 from Anthropic is now available on Vercel's AI Gateway, offering improved performance for agentic coding tasks, multi-file handling, and larger refactors. The model includes reasoning capabilities by default, enhanced vision for charts and documents, effective tool use coordination for multi-agent workflows, and support for fast mode to reduce tokens and latency.
Vercel Flags now displays live evaluation metrics on each flag's detail page, charting evaluations per minute over time with version markers to track configuration changes. You can group and filter evaluations by variant, reason, environment, SDK key, client, and more to verify rollout percentages and identify which projects use a flag.
Vercel CLI now supports inspecting feature flag version history with the new `vercel flags versions` command, allowing users to view revision history, authors, timestamps, and changed environments. A semantic diff command also shows field-level changes to targeting rules and rollout percentages.
GitHub tools are now available as an installable extension for Eve agents. Install the package, add a configuration file, and gain access to 42 GitHub tools with Vercel Connect authentication, preset scopes, and built-in approval rules for write operations.
Ant Group's Ling 3.0 Flash model is now available on Vercel's AI Gateway, offering token-efficient agentic inference with a 256K context window and free access through August 3rd. The model is optimized for high-frequency agentic workflows, coding agents, and long-context interactions without any platform fee or markup on inference.
A new Connect tab in the dashboard provides direct shell access to running Vercel Sandboxes, allowing users to run commands, manage files, and inspect ports without leaving the browser. Users can also manage sandbox lifecycle through snapshots and pause/resume controls.
97 more changes from Vercel. Sign up to read the whole changelog.
Sign up freeGitHub or email — no card needed.