---
title: "Landtest: autonomous A/B testing agent for landing pages"
description: "Landtest is an AI agent that A/B tests your landing page and learns from every result. Variants are real code in your GitHub repo, and your own host serves all traffic."
canonical: https://landtest.dev/
last-updated: 2026-10-29
---

# Landtest: autonomous A/B testing agent for landing pages

> Landtest is an AI agent that A/B tests your landing page and learns from every result. Variants are real code in your GitHub repo, and your own host serves all traffic.

## What Landtest does

- **Generates variants as real code.** Describe a change and the agent edits your landing page in your own GitHub repo, then shows you the diff.
- **Splits traffic on your own host.** Visitors are served by Vercel, Netlify or wherever you deploy. Landtest never sits in the request path, and an anti-flicker boot runs in your layout.
- **Measures results properly.** Pageviews and goals feed conversion rate, mSPRT sequential confidence and a sample ratio mismatch (SRM) check.
- **Learns.** A memory of your brand, audience, wins and analytics informs every new variant, and a Growth feed suggests what to test next.
- **Ships the winner clean.** Declaring a winner opens a change that leaves a bare page with the test runtime removed.
- **Works with your coding agent.** A hosted MCP server lets Claude, Cursor and Codex read results and run tests.

## How it works

1. Connect your repo: link your GitHub repository and select your landing page.
2. Generate AI variants: Landtest's AI creates multiple high-converting versions.
3. Run live tests: variants go live on your site with real visitors and real data.
4. Ship the winner: Landtest tells you what converts best so you can ship it permanently.
5. Learn from wins: the agent learns what worked and applies it to the next test.

## Pricing

| Plan | For | Monthly | Annual (per month) | AI credits / month | Active tests | Variants per test |
|------|-----|---------|--------------------|--------------------|--------------|-------------------|
| Starter | For developers | $25 | $20 | 100 | 5 | 2 |
| Pro | For growth teams | $50 | $40 | 200 | 20 | 5 |
| Max | For scaling startups | $200 | $150 | 1,000 | Unlimited tests | Unlimited |
| Lifetime | Bring your own API key | $299 one-time (was $599) | n/a | none (uses your own key) | Unlimited | Unlimited |

Full details: [pricing.md](https://landtest.dev/pricing.md).

## FAQ

### How does Landtest work?

Connect your GitHub repo, let Landtest generate AI variants of your landing page, and it deploys them on your existing host to split traffic and track conversions.

### Do you store my code?

Your code stays in your own GitHub repository. Variants are committed there as real, version-controlled files.

### What do credits include?

Credits cover AI variant generation. Running a test and collecting analytics does not use credits.

### How is traffic handled?

Your host (Vercel, Netlify, etc.) serves all visitor traffic directly. Landtest never sits in the request path.

### Which frameworks are supported?

Next.js (App Router) and Vite + React are supported today. Astro and plain HTML are next.

### Can I use my own domain?

Yes. Landtest works with whatever domain your site already runs on.

### Can I cancel anytime?

Yes, you can upgrade, downgrade, or cancel your plan at any time.

### Is there a limit on page views?

No limit. Track as many pageviews as your plan needs.

## For AI agents and developers

- REST API v1: https://landtest.dev/api/v1 (OpenAPI: https://landtest.dev/openapi.json; try it with the sandbox key `lt_sandbox_demo`)
- MCP server: https://mcp.landtest.dev/mcp (Streamable HTTP, OAuth 2.1 with PKCE and dynamic client registration)
- Docs MCP server: https://landtest.dev/api/docs/mcp (no auth)
- SDK and CLI: the `landtest` package and command, see https://landtest.dev/sdk
- Developer docs: https://landtest.dev/developers
- AGENTS.md: https://landtest.dev/AGENTS.md
- llms.txt: https://landtest.dev/llms.txt
- Pricing as markdown: https://landtest.dev/pricing.md
- Auth guide for agents: https://landtest.dev/auth.md