The v0 API is Vercel's headless access to v0's app-building agent: you send a prompt, v0 generates a full-stack app, runs it in a Vercel Sandbox, gives you a live preview you can embed in your own interface, and can deploy it to Vercel with one call. Vercel itself pitches it for "a white-labeled app builder: your users describe an app and get one back." This guide explains what the v0 API does in October 2026, how it is billed, what you still have to build around it, and how it compares with the Totalum API if you are launching your own AI app builder or adding one to your product. Facts are from v0.app/docs and Vercel's own announcements, read on October 18, 2026.
> Building your own AI app builder or adding one to your SaaS? The Totalum API turns a prompt into a hosted, full-stack app with its own database, login and payments, under your brand, with your users never needing a Totalum account. See the white-label AI app builder.
Quick Answer
- What is the v0 API? A REST API and TypeScript SDK (
v0) that lets your product create and iterate on apps with v0's agent:chats.createfrom a prompt,messages.sendfor follow-ups, live previews from a Vercel Sandbox, andchats.deployto ship to Vercel. - Is it new? The current API (v2) was announced as generally available in Vercel's "Introducing the new v0 API" post on August 5, 2026. An older v1 API and its docs still exist.
- How is it priced? By model tokens, drawn from your v0 credits: v0 Mini is $0.20 per million input tokens and $1.20 per million output tokens; v0 Pro is $2 and $10; v0 Max is $5 and $25. Usage is returned with each chat and message response.
- What do you still build? Your chat UI (or Vercel's starter), a server-side proxy for previews on a separate preview domain, your users' accounts, billing, and hosting decisions on Vercel.
- The alternative for a full white-label builder: the Totalum API creates each app with its own SQL database, sign-in, payments and hosting on Cloudflare, with per-project credit caps and a custom domain per project, from one API key.
What the v0 API does in 2026
The v0 API docs describe four workflows:
- Generate and iterate on apps with the Chats and Messages APIs. Each chat is an isolated workspace for one app; follow-up messages edit the app in place.
- Build a custom chat interface with
@v0-sdk/reactand the AI SDK. Vercel ships a starter:npx create-v0-sdk-app. - Embed live previews with authenticated iframe URLs. v0 runs the dev server in a Vercel Sandbox and verifies the running code, so it can catch and fix errors as it works.
- Add app generation to agents through the v0 MCP server (OAuth) or AI SDK tools (
@v0-sdk/ai-tools).
Other capabilities from Vercel's August 2026 announcement:
- Sync, async or streaming responses, with webhooks for async work. Streamed messages expose ordered "parts": text, thinking, file edits, searches, bash commands and tool calls.
- Start from existing code: create a chat from a repository, a ZIP archive or a set of files.
- Design systems as skills: load a saved design system into a request; up to three skills per request.
- Metadata to group chats by customer, workspace or app.
- Deploy to Vercel with
chats.deploy, which returns a deployment ID and a Vercel project ID.
How the v0 API is billed
v0 bills API work by model tokens, from the same credits as the v0 product (v0.app/pricing, October 18, 2026):
| Model | Input (per 1M tokens) | Cache write | Cache read | Output (per 1M tokens) |
|---|---|---|---|---|
| v0 Mini | $0.20 | $0.25 | $0.02 | $1.20 |
| v0 Pro | $2 | $2.50 | $0.20 | $10 |
| v0 Max | $5 | $6.25 | $0.50 | $25 |
| v0 Max Fast | $10 | $12.50 | $1 | $50 |
v0's plans are Free ($5 of one-time credits for new users, 7 messages a day), Plus ($30 per user per month with $30 of included credits) and Business ($100 per user per month), with extra credits on demand. Vercel's own pricing FAQ includes the question "Why do messages use so many input tokens? My prompts are short.", which tells you to plan for variable cost per customer build.
What you still have to build around the v0 API
The v0 API gives you the generation agent and a sandboxed preview. For a white-label product, Vercel's docs leave these parts to you:
- A preview proxy on a separate domain. Previews use a short-lived token that browsers cannot attach to an iframe, so you run a server route that proxies preview traffic. Because generated previews can execute untrusted code, the docs tell you to serve that proxy on a preview-only origin whose registrable domain differs from your main app's.
- Your customers' accounts and permissions, mapped to chats with metadata.
- Billing your customers for token-based usage that varies by model and by how much code the agent reads.
- Where each customer app runs:
chats.deploydeploys to Vercel; custom domains, data and per-app limits are handled on the Vercel side. - Data and backend. The API pages we read describe integrations, environment variables and MCP servers for connecting services; check how each generated app should get its database and sign-in for your use case.
None of this is a flaw. The v0 API is a powerful generation engine for teams already on Vercel. It is just not a whole builder-as-a-service.
> Want the whole pipeline managed, not just generation? Get a Totalum API key, or book a 30-minute call and we will walk you through adding an AI app builder to your product, with free credits to test.
v0 API vs Totalum API
| v0 API | Totalum API | |
|---|---|---|
| Creates full-stack apps from a prompt | Yes | Yes |
| Follow-up edits | Yes, messages to the same chat | Yes, prompts to the same project |
| Where the agent works | Vercel Sandbox, with code verification | Isolated sandbox per project; agents test the app in a real browser |
| Database per app | Not described on the API pages we read; connect services via integrations | Yes, its own SQL database per project (development and live) |
| Sign-in, payments, email | Via integrations and environment variables | Built in: BetterAuth sign-in, Stripe payments into the owner's account, email, AI, PDFs, file storage, with no keys to set up |
| Previews | Short-lived token; you run a proxy on a separate preview domain | A preview URL per project, then publish to a live URL |
| Hosting | Deploy to Vercel with chats.deploy |
Cloudflare Workers with a D1 database, no Cloudflare account needed |
| Custom domain per customer app | On Vercel | Per project via the API on any paid plan |
| Pricing model | Model tokens ($0.20 to $10 input, $1.20 to $50 output per 1M) | Credits from $0.07 each; most reads free; operation prices readable from the API |
| Budget control | Usage returned per response | Per-project monthly credit caps and live spend analytics |
| End users need a vendor account | No, your server holds the key | No, every project belongs to your account |
| Code for your customers | Chats can start from a repo; deploys to Vercel | Export or GitHub sync per project, any time |
| Starter UI | create-v0-sdk-app |
Open-source builder totalumlabs/ai-app-builder-open, one environment variable |
| Agent access | v0 MCP server, AI SDK tools | Totalum MCP server, on every plan |
Why Totalum for a white-label AI app builder
- One call, a complete app.
POST /projectswith a prompt returns a project that already has its own SQL database, sign-in, file storage and hosting. Your product shows the preview, then publishes. - Payments that go to your customer. Generated apps take payments into the app owner's Stripe account through an onboarding link, with renewals, cancellations and refunds handled automatically.
- Predictable unit economics. Credits from $0.07 each, most reads free, prices readable from the API, and per-project monthly caps so one customer cannot run away with your budget.
- Your brand, your domains. Users never see Totalum; each project can get your customer's own domain through the API on any paid plan.
- Start from a working builder. Clone the open-source reference builder, set one environment variable and rebrand it.
- Works on every plan, including Free, so you can prototype the integration before you pay.
For the architecture behind this, see our guides on how to embed an AI app builder in your SaaS via API and the white label AI app builder. If you are also weighing Lovable, our Lovable API guide explains why its REST API cannot create apps.
When the v0 API is the right choice
Pick the v0 API if:
- Your team already deploys on Vercel and wants generated apps to live there.
- You want fine control over the agent's trace (streamed parts) and are happy to build the preview proxy and the surrounding product.
- You want generation from existing repositories, or design systems loaded as skills.
- You are adding app generation to an agent built on the AI SDK.
Pick the Totalum API if you want the full builder-as-a-service: database, sign-in, payments, hosting, domains and per-customer budgets handled for you. If you only need v0 as an end-user tool, our v0 alternatives roundup compares it with other builders.
FAQ
Does v0 have an API?
Yes. The v0 API gives your product v0's app-building agent: create apps from prompts, iterate with follow-up messages, embed live previews and deploy to Vercel. The current version was announced as generally available on August 5, 2026.
How much does the v0 API cost?
It is billed by model tokens from your v0 credits: v0 Mini costs $0.20 per million input tokens and $1.20 per million output tokens, v0 Pro $2 and $10, and v0 Max $5 and $25. Plans range from Free to $100 per user per month.
Can I build a white-label app builder with the v0 API?
Yes. Vercel lists "a white-labeled app builder" as a use case. You build the chat interface (or start from create-v0-sdk-app), a preview proxy on a separate domain, your user accounts and billing, and you deploy generated apps to Vercel.
What is the difference between the v0 API v1 and v2?
v2 is the current, generally available API announced in August 2026, built around chats and messages with Sandbox previews, streaming, webhooks and one-call deploys. The v1 docs remain available for existing integrations.
Is there an alternative to the v0 API?
Yes. The Totalum API creates complete apps with their own SQL database, sign-in, payments and hosting from one call, bills in credits from $0.07 each with per-project caps, gives each project a custom domain on paid plans, and works on every plan, including Free.
Do my users need a v0 or Totalum account?
No, in both cases. Your server holds the API key, and the apps belong to your account.
Launch your own AI app builder
The v0 API is a strong generation engine, especially for teams on Vercel. If you would rather ship a complete AI app builder, with databases, sign-in, payments, hosting and domains handled for every customer, get your Totalum API key, read the API docs, or book a call.