Vercel is a cloud platform for building, deploying, and scaling web applications and AI features, combining a global edge network, serverless and fluid compute, managed storage, security, observability, and integrated AI infrastructure.
Software capabilities AI (Gateway, SDK, Agent & v0) AI Agent Observability Traces AI agent runs and exports AI SDK spans with OpenTelemetry, showing a trace hierarchy in the dashboard. AI Gateway App Attribution Attributes gateway usage and spend to specific apps and users through request headers. AI Gateway Automatic Caching Automatically caches prompt prefixes to reduce token cost and latency on repeated requests. AI Gateway Batch Processing Submits up to 1,000 text generation requests as a single asynchronous batch job for OpenAI and Anthropic models at 50% of standard token prices, with regional inference pinning for Anthropic models; batches reach a terminal state within 24 hours. AI Gateway Bring Your Own Key Routes requests using a customer's own provider API keys, including request-scoped keys, through the gateway. AI Gateway Embeddings and Reranking Generates vector embeddings for semantic search and re-scores candidate documents by relevance to improve retrieval-augmented generation pipelines. AI Gateway Evaluation Evaluates shared state against typed questions through AI Gateway and returns structured answers: boolean probabilities, a choice from a named set, or a score along an ordered scale, with many questions per request. AI Gateway Fast Mode Routes requests to a faster serving path for supported models across providers through the unified gateway.speed option or an explicit fast-slug model ID, with automatic fallback to the base model and speed-dependent billing. AI Gateway GPT-Live Serves OpenAI GPT-Live full-duplex voice sessions through AI Gateway over a dedicated WebSocket endpoint, with client delegation, context appends, and per-second duration billing. AI Gateway Image Generation Generates and edits images from text prompts across multiple multimodal and image-only models through a unified API, using the AI SDK's stable generateImage function (promoted from experimental_generateImage). AI Gateway Inputs and Tools Sends images, PDFs, audio and video to models through AI Gateway and wires tool calling, with per-format examples for AI SDK, Chat Completions, Messages API and Responses. AI Gateway Leaderboards Ranks models, model labs, opted-in apps, and inference providers by anonymized daily usage share observed through AI Gateway, with shareable chart images and an open CSV data export. AI Gateway Model Fallbacks Automatically retries a failed request against fallback models and providers to improve reliability. AI Gateway Model Routing Routes requests to hundreds of AI models across providers through one API key, with configurable provider ordering, filtering, timeouts, and automatic fallback on failure. AI Gateway Observability and Spend Logs latency, token usage, and spend for every request, with a dedicated Logs page (live tail, per-request fallback-path and cost detail, filtering, CSV/JSON export) and Trace Drains that forward an OpenTelemetry trace of every request to OTLP endpoints or native integrations (Braintrust, Dash0, Kubiks, Sentry, Statsig). AI Gateway Private Inference Routes inference to private or region-pinned providers with data residency and regional routing controls, over endpoints speaking the OpenAI Chat Completions, OpenAI Responses, Open Responses, or Anthropic Messages API format. AI Gateway Provider Routing Orders, filters, and sorts providers per request and applies routing rules to rewrite or deny models. AI Gateway Reasoning Passes through provider reasoning and thinking configuration and returns reasoning output across supported models, including a unified top-level reasoning effort option (AI SDK 7) mapped to each provider's native config, and cross-format translation so a reasoning parameter set in any supported API format works against any reasoning-capable model. AI Gateway Responses Compaction Compresses a long OpenAI Responses API conversation into a single compaction item through POST /v1/responses/compact, called automatically by coding agents such as Codex. AI Gateway Safety Identifiers Lets requests carry a per-end-user safety identifier that AI Gateway forwards to the serving provider (OpenAI safety_identifier or Anthropic metadata.user_id) for per-user abuse detection, surviving provider and model fallbacks. AI Gateway Security and Compliance Enforces zero data retention, blocks prompt training, requires HIPAA-compliant providers, restricts requests to region-pinned or allowlisted providers and models, and supports bring-your-own-key credentials, including marking a BYOK key as ZDR-compliant; each team-level toggle is recorded in the activity log. AI Gateway Service Tiers Lets requests to OpenAI, Google AI Studio, and Google Vertex AI models specify a service tier (default, priority, or flex) that trades processing priority and latency against cost, falling back to the default tier on a best-effort basis when a provider does not grant the requested tier. AI Gateway Speech and Realtime Voice Converts text to spoken audio, transcribes audio to text (including streaming transcription over WebSockets with short-lived browser client tokens), and supports low-latency speech-to-speech realtime voice agent sessions over WebSockets. AI Gateway Stripe Billing Meters and bills end-user AI usage through a built-in Stripe billing integration with metered events. AI Gateway Text Generation Generates and streams text completions, tool calls, and structured output from language models through a unified API across providers. AI Gateway Video Generation Generates, edits, and extends videos from text prompts, images, or reference video across multiple video-generation providers. AI Gateway Virtual Models Defines virtual models that bundle routing, provider, and fallback settings under a single reusable model name, managed through a REST API that reads, updates and soft-deletes a config by its slug and lists configs with cursor pagination. AI Gateway Web Search Tools Lets AI models call built-in search tools, including Perplexity, Exa, Parallel Search, Tako Search over a curated knowledge graph, or provider-native search, from the AI SDK or as server tools on the OpenAI-compatible Chat Completions API. AI SDK Core Generation Provides generateText, streamText, generateObject, and streamObject functions for producing free-form or schema-validated output from any supported model with a two-line provider switch. AI SDK for Python Provides a Python toolkit (pip/uv install ai, Python 3.12+) for building LLM apps and agents that routes through AI Gateway by default, with streaming responses, Pydantic-validated structured outputs, and an @ai.tool/ai.Agent tool-calling loop. Public beta. AI SDK Framework Ecosystem Integrates with React, Vue, Svelte, and Node.js applications and interoperates with third-party agent frameworks through Chat Completions-compatible or native adapters. AI SDK Tool Calling Lets a model call developer-defined functions during generation and use their results to produce a final answer, including via ToolLoopAgent, a reusable agent class that runs a tool-calling loop instead of hand-rolling one around generateText. eve Agent Framework Provides an agent framework for defining durable agent projects with sessions, tools, skills, subagents, and connections. Responses API over WebSocket Keeps one WebSocket connection open across multiple turns of the OpenAI Responses API through AI Gateway, removing the per-turn connection handshake to cut latency in agent loops that make many requests. v0 App Generation Generates and iterates on React and Next.js UI and application code from natural-language prompts, producing deployable projects. Vercel Agent Chat Answers natural-language questions about a team's projects, deployments, and usage in a dashboard chat interface, grounded in real project data. Vercel Agent Code Review Automatically reviews pull requests, generates patches, and validates fixes by running a project's real builds, tests, and linters inside a secure sandbox before suggesting them. Vercel Agent Installation Analyzes a repository and opens a pull request that installs and wires up Web Analytics or Speed Insights without manual integration work. Vercel Agent Investigation Automatically investigates production anomaly alerts by querying logs and metrics around the time of the alert and summarizing likely root causes.
Deployments & Builds .vercelignore Excludes specified files and directories from a deployment via a .vercelignore file in the project root, with an optional allowlist mode that ignores everything except explicitly allowed paths. Build Cache Caches build artifacts and dependencies between deployments to speed up subsequent builds. Build Machines Runs builds on Basic, standard, enhanced, turbo, or elastic machines with selectable size and resources; Basic (2 vCPU, 8 GB) is included on Hobby and billed at $0.007/build-minute on Pro and Enterprise. Build Pipeline Builds projects using auto-detected framework settings with configurable install, build, output, and ignore commands and a root directory. Claim Deployments Transfers ownership of an anonymously created deployment to a user or team account through a claim flow. Code Owners & Approvers Requires designated owners or approvers to review changes to matched file paths before they can be merged. Concurrent Builds Runs multiple builds in parallel with on-demand concurrency, prioritizing production builds and queueing the rest. Conformance Enforces code-quality, performance, and security rules across a monorepo with configurable rules and allowlists. Custom Environments Defines named custom environments such as staging in addition to the default production and preview environments. Deploy Button Provides an embeddable button that clones a template repository and deploys it with prompted environment variables and integrations. Deploy Hooks Triggers a deployment by sending an HTTP request to a project-specific hook URL without a git push. Deployment Bisect Binary-searches the deployment history, optionally with a test script, to find which deployment introduced a regression. Deployment Checks Runs automated checks against a deployment, such as performance or integration tests, that can block promotion to production. Deployment Retention Sets policies for how long deployments are kept before automatic deletion, with restoration of deleted deployments. Flags SDK Provides a framework-native library for Next.js and SvelteKit to define feature flags in code with TypeScript support, working with Vercel Flags or a connected Marketplace provider. Git Deployments Deploys automatically on every push to connected GitHub, GitLab, Bitbucket, or Azure DevOps repositories. Instant Rollback Reverts production to a previous deployment instantly without rebuilding, with the option to undo the rollback. Monorepo Support Deploys multiple projects from a single monorepo with filtered installs and skipped builds for unaffected projects. Preview Deployments Creates an isolated deployment with a unique generated URL for every branch and commit before it reaches production. Production Deployments Promotes builds from the production branch to production and assigns production domains and aliases. Project Configuration Configures projects declaratively through a vercel.json or vercel.toml file, or programmatically through a vercel.ts file. Remote Caching Shares Turborepo and Nx build cache artifacts across machines and external CI through the Vercel Remote Cache; external CI/CD can authenticate via OpenID Connect (recommended) or a Personal Access Token. Rolling Releases Gradually shifts production traffic to a new deployment across configurable stages, with observability and the ability to advance, abort, or complete the rollout. Skew Protection Pins each client's requests to the deployment version it originally loaded to prevent version mismatch errors during rollouts. Vercel CI Runs Turborepo tasks in isolated compute environments using the task graph and Remote Cache to skip unaffected work, with per-task CPU allocation and CI-scoped environment variables. Vercel Drop Deploys a file or folder by dragging it into the browser without configuring a git repository. Vercel Flags Provides a built-in feature-flag platform with entity-based targeting rules over string, number, boolean, string-array and timestamp attributes, weighted splits, and progressive rollouts, plus a Flags Explorer for overriding flags in the browser.
Edge Network, Rendering & Domains Cache Invalidation Purges CDN and data cache entries — including a dedicated path for immutable static assets and vendor-added system cache tags (e.g. Image Optimization's source-image tag) — by path, tag, or full deployment through the API or dashboard. CDN Cache Caches static and dynamic responses at the CDN using Cache-Control directives such as s-maxage, stale-while-revalidate, and stale-if-error. CDN Compression Automatically compresses responses with gzip and Brotli based on the client's accepted encodings. Custom Error Pages Serves custom pages for platform and application errors using error tokens. Custom Headers Sets custom response headers and CORS rules on matched routes. Custom SSL Certificates Uploads and pre-generates custom SSL certificates in addition to automatically provisioned ones. Data Cache Caches data-fetch results with tag-based and time-based revalidation for Next.js applications, sharing storage across a team's projects on Hobby and Pro and isolating storage per project on Enterprise. Directory Listing Displays the contents of a directory when a user visits its path if no index file is present, toggleable per project and suppressible for specific directories. Domain Registration & Transfer Registers, purchases, transfers, and renews domain names directly through Vercel with auto-renewal and redemption handling. Domains and DNS Registers, connects, and manages custom domains, DNS records, nameservers, and automatic SSL certificate provisioning for deployments. Draft Mode Renders draft or unpublished content in preview by bypassing static caching for editors. Global Config Provides a low-latency global key-value data store — renamed from Edge Config to Global Config — for reading configuration such as feature flags, redirects, or IP blocklists close to the requesting user; the CLI keeps `vercel edge-config` as a backward-compatible alias for the new `vercel global-config` command. Image Optimization Dynamically resizes, compresses, and converts images to modern formats on request and caches the transformed results on the CDN. Incremental Static Regeneration Serves cached page responses under a stale-while-revalidate pattern and regenerates them in the background on a time interval or on-demand invalidation. Open Graph Image Generation Generates dynamic Open Graph social images at request time from HTML and CSS with the @vercel/og library. Partial Prerendering Combines a cached static shell with dynamically rendered content streamed into the same response, personalizing pages without sacrificing time-to-first-byte. Project-Level Routing Rules Defines project-level routing rules for rewrites, redirects, headers, and blocks that are staged and published independently of deployments. Redirects Configures URL redirects at deployment, project, or bulk scale with versioning, staging, and CSV or JSON upload. Request Collapsing Collapses concurrent identical origin requests into a single upstream fetch to reduce load. Rewrites Rewrites request paths to same-application routes, external origins, or internal services. Routing Middleware Runs code before a request reaches the application to implement custom routing, rewrites, redirects, and header rules, from a middleware file at the project root or from any entrypoint and matcher declared in the vercel.json proxy property. Runtime Cache Provides a framework-agnostic runtime cache API for caching arbitrary values across function invocations, isolating storage per project on Pro and Enterprise and sharing storage across a team's projects on Hobby. Vercel CDN Serves and caches deployments from a globally distributed network of points of presence and compute regions, applying routing, compression, and security before requests reach application code.
CLI, API & Platform Access Tokens Creates and manages scoped API access tokens for authenticating to the Vercel API and CLI. Agent Skills Directory Publishes an installable directory of packaged agent skills that extend AI coding assistants with specific procedural capabilities, installed through the skills CLI. Build Output API Defines a framework-agnostic build output format that any framework can produce to run on Vercel, including a services property for declaring multiple independently built services with their own output tree and routable bindings between them. Checks API Lets integrations post status checks and rich results that can gate deployment promotion. Integration Agent Tools Lets Vercel Agent and AI agents call integration tools to read and act on connected resources with optional read-only mode. Key Management (KMS) Signs JWTs and messages from Vercel Functions with Vercel-managed keys, authenticated by the function's OIDC token, with issuers, key rotation, and project and Connect grants. Markdown & Agent Discovery Serves documentation to agents as .md endpoints plus a cross-link graph (graph.json and per-page .graph.md) and site-wide discovery files, honoring an explicit Accept: text/html or text/markdown preference before falling back to agent detection. MCP Design Import Imports a self-contained HTML bundle from Claude Design via a public claudeusercontent.com URL and deploys it to Vercel through an MCP tool. MCP Server Hosting Hosts Model Context Protocol servers on Vercel Functions with fluid compute concurrency and built-in OAuth authorization for AI tool access. OAuth Apps Registers OAuth applications that authenticate against the Vercel API on a user's behalf. OIDC Federation Issues OIDC tokens so functions can access AWS, GCP, and Azure and customer APIs without static long-lived secrets. Platform Elements Provides prebuilt React UI blocks and server actions such as custom-domain, DNS table, deploy popover, deploy files and report-abuse for building platforms, installed with the Vercel Platforms CLI or from a hosted shadcn registry. Private Package Registry Distributes Vercel-internal npm packages under the @vercel-private scope through a private registry that requires account authentication and access to the corresponding Vercel product. Sign in with Vercel Lets third-party applications authenticate users with their Vercel account through an OAuth 2.0 and OpenID Connect identity provider operated by Vercel. Vercel CLI Provides a command-line interface for deploying, linking, configuring, and inspecting projects, deployments, domains, resources, and trace sampling rules. Vercel for Platforms Provides templates and primitives for building multi-tenant SaaS platforms with per-customer subdomains, custom domains, and isolated deployments. Vercel MCP Server Exposes Vercel's own hosted MCP server so AI clients can search docs, manage projects/deployments/domains, query Web Analytics, inspect runtime/build logs and grouped runtime error clusters, and deploy file trees directly, through tools such as search_vercel_documentation, get_web_analytics, get_runtime_errors, and deploy_to_vercel. Vercel Plugin for AI Coding Agents Installs Vercel-specific context, skills, specialist agents, and slash commands into supported AI coding tools such as Claude Code, OpenAI Codex, Cursor, GitHub Copilot, and Grok Build. Vercel REST API Exposes a REST API for programmatic control of deployments, projects, domains, DNS, environment variables, and resources. Vercel SDK Provides a type-safe TypeScript client library for calling the Vercel REST API without hand-writing HTTP requests. Webhooks Delivers account and integration webhook events for platform activity with signature verification and retries.
Firewall & Application Security Attack Challenge Mode Challenges all visitors with a browser verification during an active attack to filter automated traffic. Bot Management Manages bot traffic with managed bot-protection and AI-bot rulesets and verified-bot handling. BotID Runs an invisible client-side machine-learning challenge that distinguishes human visitors from automated bots on high-value routes such as checkout and signup. Compliance Certifications Maintains SOC 2 Type 2, ISO 27001, PCI DSS, EU-U.S. DPF, and HIPAA BAA attestations with a shared responsibility model; compliance documents can be downloaded individually or as a bulk zip from the dashboard. Content Security Policy Configures Content Security Policy and other security headers for deployments. DDoS Mitigation Automatically detects and mitigates distributed denial-of-service attacks at the network edge. Deployment Protection Restricts access to preview and production deployments through Vercel Authentication, Password Protection, Passport (external identity provider), or Trusted IP allowlisting. Encryption and TLS Encrypts traffic with modern TLS, HSTS, OCSP stapling, and post-quantum ciphers, managing certificates automatically. IP Blocking Blocks or bypasses traffic by IP address or CIDR range at project and account level. Page Integrity Detects and blocks client-side script tampering and skimming attacks with alerting and blocking modes, plus script approval rules to allow or remove specific scripts from the inventory. Passport Restricts deployment access through an external identity provider configured via a Vercel Connect application, and exposes verified visitor identity to the application and downstream backends via the @vercel/passport SDK. Generally available on the Enterprise plan (previously beta). Protected Source Maps Restricts access to production source maps to authenticated team members. Protection Bypass for Automation Lets automated systems and CI bypass deployment protection using a shared secret. Rate Limiting Rate-limits requests by custom keys through firewall rules and the @vercel/firewall SDK. Secure Compute and Static IPs Connects Vercel Functions to private backend infrastructure through dedicated VPCs with VPC peering, shared static egress IPs, or AWS PrivateLink connections. Security Dashboard Runs a team's security checks and reports their status in a dashboard, with the same checks available from the vercel security CLI command. Shareable Links Generates time-limited links that grant external users access to protected deployments. Trusted Sources Restricts which Vercel projects and external services are allowed to reach a deployment. Vercel Firewall and WAF Inspects incoming traffic and lets teams log, challenge, deny, or rate-limit requests using automatic DDoS mitigation, custom rules, and managed rulesets such as OWASP. WAF Custom Rules Defines custom firewall rules with conditions and allow, deny, challenge, or rate-limit actions, staged and published separately from deployments. WAF Managed Rulesets Applies managed rulesets including the OWASP core ruleset, bot protection, and AI bots.
Functions & Compute Backend Frameworks Deploys Node.js and Python backend frameworks such as Express, Fastify, Hono, NestJS, FastAPI, Flask, and Django as functions with zero configuration. Container Image Functions Runs functions from custom OCI container images with port resolution and scale-in behavior. Cron Jobs Triggers a Vercel Function on a schedule defined by a cron expression in the project configuration; the CLI propagates top-level schedules from vercel.json into the Build Output API's config.json during vercel build and validates it against the schedules schema. Edge Runtime Runs lightweight functions on a V8-based Edge Runtime with a Web-standard API close to users. Environment Variables Manages project environment variables scoped to Development, Preview, and Production, with a write-only Secret type for credentials and readable Config values. Fluid Compute Serves many concurrent requests from a single function instance with active-CPU billing, bytecode caching, and cold-start prevention. Function Configuration Configures function memory (up to 4 GB), maxDuration (up to 800s GA, 1,800s beta on supported Node.js, Bun, and Python runtimes), and per-route settings in code or vercel.json. Function Runtimes Runs functions in Node.js, Python, Go, Ruby, or Rust official runtimes as well as community runtimes. Python functions are automatically compiled to bytecode (.pyc) during builds when space allows, reducing cold-start initialization time. Function Streaming Streams function responses incrementally to the client as they are generated. Large Functions Raises the function package size limit to 5 GB (beta) for Node.js, Bun, and Python deployments on Fluid compute, above the 250 MB standard limit. Microfrontends Splits a single application into independently deployable microfrontend projects that Vercel routes and serves together on the network as one cohesive site. Multi-Region Functions & Failover Deploys functions to one or more regions with automatic regional failover. OpenAI Agents API Integration Connects OpenAI Agents API sessions to Vercel Sandbox for isolated code execution and persistent workspace files. Services Deploys multiple backend and frontend services within a single Vercel project with shared public routing and private internal service bindings for server-to-server calls; multi-service build output is prefixed per line with a [vc:service:<name>] tag the build container uses to attribute log lines to a service. Vercel Container Registry Stores, tags, and manages container images in a Vercel-hosted registry for use by functions and sandboxes, with per-repository team permission grants and a bulk action to clear all team permissions on a repository. Vercel Functions Runs application code as auto-scaling serverless functions with a configurable runtime (Node.js, Python, Go, Ruby, or Edge), region, memory, and duration. Vercel Queues Provides a durable, at-least-once message queue with topics and independent consumer groups for decoupling request handling from asynchronous background processing, including a Celery adapter, a poll-mode broker for workers running outside Vercel, and replay of a topic's non-expired history (e.g. giving each viewer an independent replay of an agent's file edits, terminal commands, and tool calls). Vercel Sandbox Provisions isolated, persistent Linux microVMs for running untrusted or AI-generated code from a selectable image (default vercel/sandbox/universal, 64 GB ephemeral NVMe storage on SDK 3.0+), with snapshotting, multi-user isolation, session expiry, usage reporting, configurable default and failover regions per project, and forking that inherits the source sandbox's state. Vercel Schedules Creates and manages recurring (cron) or one-time schedules that deliver payloads to queue topics, with jitter, timezone support, enable/disable, and immediate invocation. Vercel Workflows Runs durable, multi-step functions written with workflow and step directives, handling retries, long sleeps, and waiting on external events automatically. WebSockets Serves persistent WebSocket connections from Vercel Functions, with framework guides for Node.js/Nitro/Nuxt (native), Python FastAPI (ASGI), Django (via Django Channels), and Flask (via flask-sock).
Observability & Analytics Activity Log Records team and project activity events for review in the dashboard and CLI. Alerts Configures built-in and custom alert rules with Slack and webhook notifications and optional AI investigation of anomalies, filterable by rule type from the CLI. Build Logs Provides build-time logs for inspecting and debugging deployments. Custom Event Tracking Tracks custom client-side and server-side analytics events with custom data and dashboard filtering. Drains Streams logs and telemetry from deployments to external HTTPS endpoints, with payloads covering request, proxy, firewall and custom environment fields. Monitoring Queries and charts metrics across projects with a custom query builder over request and traffic data. Notebooks Collects and organizes saved observability queries into shareable notebooks with line-chart, volume-chart, table, or big-number views for reuse and team collaboration. Observability Provides dashboards and a query interface over functions, middleware, edge requests, caching, and traffic to debug and monitor deployed applications. Observability Plus Adds extended data retention, advanced metrics, a query engine, and alerts to Observability as a paid add-on. OpenTelemetry Instrumentation Instruments applications with OpenTelemetry through a Vercel instrumentation hook and custom spans. Runtime Logs Streams and searches runtime function and edge logs with field filters, retention windows, and sharing. Session Tracing Runs on-demand session and page traces from the dashboard or CLI to debug end-to-end performance. Speed Insights Measures real-user Core Web Vitals and related performance metrics for deployed pages and scores them against real-world benchmark data. Tracing Captures OpenTelemetry traces of requests with automatic instrumentation and a dashboard trace viewer. Web Analytics Tracks page views, unique visitors, referrers, and custom events for deployed applications without using cookies to identify visitors.
Access & Permissions Access Groups Groups members with defined project access that can be synced from an external directory. Account Management Manages a user's account profile, email addresses, and connected identities. Audit Logs Records team member actions and security-relevant events for export as CSV or streaming to external storage for compliance review. Enterprise Managed Users Provisions fully managed enterprise user accounts controlled by the organization's identity provider. Extended Permissions Grants granular extended permissions and project roles beyond the base team roles, including permissions for AI Gateway settings, budgets and credits, connectors, environments and environment variables, integrations, project creation, production deployments, usage viewing, v0 access, and workflow decryption, with several roles and permissions assignable to one member at once. Login Methods Supports email and Git-provider login through GitHub, GitLab, Bitbucket, and Google. Organizations Groups multiple teams under a single organization for centralized management and billing, using an enterprise billing plan for internal teams or a platform billing plan for up to 100 multi-tenant child teams, with a REST API for creating and managing organizations and their teams. Protected Git Scopes Restricts which Vercel teams or accounts may deploy from an organization's Git repositories, configurable per team with up to five protected scopes. Role-Based Access Control Assigns team-level and project-level roles and access groups that determine which resources and actions each team member can view or change. SAML SSO and Directory Sync Authenticates team members through an organization's SAML identity provider and syncs team membership from an external directory via SCIM. Teams Organizes members, projects, and billing under shared teams with collaboration settings. Two-Factor Authentication Requires two-factor authentication for accounts with optional team-wide enforcement.
Billing & Administration Account Plans Offers Hobby, Pro, and Enterprise plans with different features, limits, and trials. AI Credits Manages prepaid AI Gateway and Vercel Agent credit balances with top-up and auto top-up. AI Gateway Invoiced Billing Lets Enterprise teams pay for AI Gateway usage by invoice instead of prepaid credits, with no payment processing fees, set up through Sales. AI Gateway Volume Discounts Discounts AI Gateway token spend two ways: models priced below the provider's list price for every team, and negotiated custom discounts for volume commitments starting at six figures, alongside zero-fee ACH payment for volume customers. Billing and Invoices Manages payment methods, invoices, partial invoices, taxes, and custom invoicing. Flat Rate CDN Offers fixed monthly CDN pricing across capacity tiers covering CDN requests and data transfer instead of per-unit billing, with per-project exclusion onto the Flex CDN network. Flexible Commitment Offers enterprise flexible spend commitments applied across usage-based products with regional pricing. Spend Management Sets spend limits, alert thresholds, and automatic project pausing to control on-demand costs. Support Center Submits and manages support tickets with response levels that vary by plan. Team Seats Manages developer, viewer, and billing seats and their per-seat charges. Usage Dashboard Shows resource usage and cost breakdowns per resource in a dashboard with viewing options, including Deployment Storage and Functions Storage ($0.10/GB-month each) for retained deployment output. Usage-Based Billing Meters and bills usage across compute, network, storage, and other resources with included credits and regional pricing.
Collaboration & Toolbar Accessibility Audit Tool Audits pages for accessibility issues and records them from the Vercel Toolbar. Edit Mode Highlights CMS-backed content elements on a live page as clickable Content Links from the Vercel Toolbar so authors can jump straight to the matching CMS field without a developer. Flags Explorer Views and overrides feature flags in the browser from the Vercel Toolbar and shares overrides. Interaction Timing Tool Measures interaction-to-next-paint timing from the Vercel Toolbar. Layout Shift Tool Visualizes and inspects cumulative layout shift on a page from the Vercel Toolbar. Notifications Delivers deployment, comment, and account notifications across web, email, and integrations. Preview Sharing Shares preview deployments with team members and external collaborators, including under deployment protection. Vercel Toolbar and Comments Adds an in-browser toolbar to preview deployments that lets collaborators leave comments, preview draft content, inspect feature flags, and audit performance.
Marketplace & Integrations Build Integrations Provides REST APIs, flows, and secrets rotation for partners to build and list native and Marketplace integrations. CMS Integrations Connects headless CMS platforms to Vercel projects for environment-variable import, in-browser edit mode, and draft mode previews of unpublished content. Connectable Accounts Connects existing external provider accounts to Vercel projects without provisioning new resources. Git Integrations Integrates with GitHub, GitLab, Bitbucket, Azure DevOps, and Origin (Cursor's Git platform) for deployment automation, commit statuses, and pull-request comments. Integration Billing & Monetization Lets Marketplace partners meter usage and bill customers through Vercel's unified billing and invoicing via the Partner API. Native Integrations Provisions native integration products such as databases and AI providers with automatically configured, prefixed environment variables. Vercel Connect Brokers authenticated connections from Vercel to third-party provider APIs (Slack, GitHub, Linear, and others) through managed connectors, tokens, and triggers, with REST endpoints for per-project connector association, trigger destinations, and framework adapters including an eve entrypoint (@vercel/connect/eve). Vercel Marketplace Provisions and manages third-party databases, AI providers, feature-flag platforms, and other resources directly from the Vercel dashboard with unified billing.
Storage Blob Multipart Uploads Uploads large files in parts through client-side or server-side multipart uploads with progress and abort support. Blob Signed URLs Issues signed tokens and presigned upload and download URLs for controlled blob access. Marketplace Storage Provisions third-party Postgres, Redis, and vector databases from the Marketplace with connection strings and unified billing, and lets teams run Redis commands and SQL queries directly against provisioned resources from the dashboard. Private Blob Storage Stores private blobs delivered only through signed URLs or token-based access control. Public Blob Storage Stores public blobs served over a direct URL without authentication, for images, videos, and other public assets. Vercel Blob Provides object storage for uploading and delivering images, videos, and other files from public or private stores, with client and server upload paths. Vercel Drives Provides persistent, attachable object storage with a lifecycle independent from any single sandbox, defaulting to 1 TiB with a 16 TiB per-drive quota, mounted at sandbox creation or changed later through the sandbox update API.
Vercel products and featuresvercel.com