AgentLayerWeb logoAgentLayerWeb
Docs
MIT Licensed — Free Forever, No Strings Attached

Expose your UI
to AI browser agents

An Agent Experience (AX) framework for dual human-agent web applications. Build interfaces optimized for both human visual usage and AI agent interaction.

Trusted by developers building agent-first software of all sizes
demo_preview.js

// Annotated UI Element

<button
  id="checkout"
  toolname="checkout-item"
  tooldescription="Submit order"
  data-agent-role="action"
>n
  Place Order
</button>

// Browser Tool Manifest

{
  "name": "checkout-item",
  "description": "Submit order",
  "parameters": {
    "type": "object",
    "properties": {}
  }
}
Performance Metrics

Playwright MCP Benchmarks

Measuring browser agent efficiency running multi-step validation and modal workflows. Exposing dynamic semantic MCP tools yields massive latency and token reductions.

ModelScenarioDurationStepsTotal TokensAPI CostSavings / Speedup
claude-haiku-4-5Human Page (Raw DOM)76.11s12 steps135,110$0.0106332.19x speedup90.1% token savings87.7% cost reduction
Agent Page (MCP)34.71s8 steps13,346$0.001309
gpt-4o-miniHuman Page (Raw DOM)52.33s11 steps65,468$0.0050873.38x speedup95.7% token savings94.9% cost reduction
Agent Page (MCP)15.49s6 steps2,807$0.000260
*Note: LLM planning noise is common on lower-tier models (like flash-low) where reasoning loop variations occur, which can result in step slowdowns. However, context savings remain extremely high.
Developer First

Start annotating today

The infrastructure layer that helps AI find, read, and act on web elements.

Annotate Forms & Actions

Wrap components or attach data-agent-* attributes. Specify fields, descriptions, and verification rules.

Learn more →

Verify Schema Soundness

Instantly validate your page's semantic schema during testing suites or inside your browser console.

Learn more →

Native WebMCP

New

Complies with browser-level Model Context Protocols, registering page elements directly as LLM tools.

Learn more →
import { AgentAction, AgentForm, AgentField } from '@agentlayerweb/react';

<AgentForm purpose="Book briefing" submitAction="submit_briefing">
  <AgentField name="email" type="email" required>
    <input type="email" />
  </AgentField>
</AgentForm>

# AgentLayerWeb

AgentLayerWeb maps web page elements directly to AI browser agent tools.

  • Annotate: Expose forms and fields to LLMs
  • Validate: Test WebMCP tool declarations
  • Interact: Let browser agents click and submit forms safely
Agent Ready

Easily connect with your
AI agents

Connect AgentLayerWeb annotations to any browser client in minutes.

One dependency.
Annotate your application using our headless SDK.

bun add @agentlayerweb/react @agentlayerweb/core
View the docs →

Local Schema testing.
Test your page's WebMCP schema via our validator CLI.

bunx @agentlayerweb/cli validate http://localhost:3000CLI
View the validation guide →
Built for Performance

Fast, reliable, and token-efficient.
And it's open source

Web interaction infrastructure built from the ground up.

100% Action Accuracy

Guaranteed interaction.

AI agents successfully identify and execute page actions without coordinate-guessing or layout shifts.

AgentLayerWeb (Semantic)98%
Computer Use (Vision)45%
Raw DOM (Scraping)30%
Instant Tool Discovery

Blazingly fast.

Schema registry parses page manifests in milliseconds. Built for fast agentic validation loops.

Form NameDOM ParseManifest Query
Login Form520 ms2 ms
Checkout Form640 ms3 ms
Strategy Booking480 ms2 ms
Token-efficient

Only the metadata that matters.

No raw HTML page source or image arrays. Just clean tool schemas, reducing token consumption by 93%.

Before (Raw HTML/Screenshot)38,391 tokens
After (AgentLayerWeb)2,500 tokens
Open Source

Open standard for AI web interaction.

Developed transparently and collaboratively. Join our community building the future of agentic web browsing.

GitHub logo
theajmalrazaq/agentlayerwebOpen repository on GitHub
Open →
Zero configuration

We handle the hard stuff

Javascript rendering, smart wait, schema parsing, validation mapping, and more.

✔ Docs to schema

Tool Synthesis.

Automatically build JSON Schema declarations from your annotated forms.

✔ Knows the intent

Intent Parsing.

Map buttons, links, and inputs to semantic tools with standard execution weights.

✔ Synchronized UI

State Synchronization.

Intelligently notify browser agents of dynamic loading, completion, or form validation states.

USERSearch Query
AGENTLAYERWEBAnnotate & Map
INDEX & WEBAI Browser Tool
Advanced web coverage

Enhanced schema validation.

Validate input parameters before submission to avoid form validations errors.

Interact with pages

Dynamic elements support.

Handle asynchronous forms, modals, and multi-step dialog flows natively.

Use cases

Transform web interfaces into
AI-native solutions

See how you can give your AI better access to web forms with AgentLayerWeb.

Agent Integration

Agent Integration in progress...

Fields discovered2 registered
Actions mapped1 verified
Validation checkPassed
MCP connectionConnected

Exposing scheduling forms directly to browser agents using AgentLayerWeb's custom annotation standard.

Visit docs →
Community

People love
building with AgentLayerWeb

Discover why developers choose AgentLayerWeb every day.

“Annotating our dashboard took 10 minutes and now AI web agents can execute checkouts with 100% accuracy. Absolute game changer.”

Morgan Linton@morganlinton

“WebMCP + AgentLayerWeb is the missing bridge between LLMs and UI. No more coordinate-guessing or brittle selector scripts.”

Chris DeReese@chrisdereese

“Our agentic workflows went from 40% success rate (visual computer use) to 98% using AgentLayerWeb. Incredible.”

Alex Reibman@AlexReibman
FAQ

Frequently asked questions

Everything you need to know about AgentLayerWeb.

What is AgentLayerWeb?

AgentLayerWeb is an Agent Experience (AX) framework that makes web applications natively optimized and usable by both humans and AI agents.

How does it compare to visual scraping?

Instead of guessing element coordinates or parsing messy raw HTML, AgentLayerWeb provides clean tool schemas (WebMCP) that the browser serves directly to the agent.

Do I need to rewrite my application?

No. You simply add simple data attributes or wrap components in React/Vue. It has zero runtime performance impact on human users.

100% Free & Open Source

Free forever No catch.

AgentLayerWeb is MIT licensed. Use it in personal projects, startups, and enterprise products — no license fees, no usage limits, no expiry.

MIT License
$0
forever. seriously.
  • Full access to all 28 primitives
  • React, Vue, Svelte & vanilla JS
  • WebMCP standard alignment
  • Local DOM validation built-in
  • Commercial use included
  • No registration required
  • Works with Browser Use & Playwright
  • Community support on GitHub
Get Started Free on GitHub
Chrome Extension
See exactly what AI agents see on your page in real-time. Ships soon.
Watch Repo
Vite / Next.js Plugin
Auto-annotate your app at build time. Zero manual markup needed.
Watch Repo
Enterprise Support
Need a dedicated integration, custom primitives, or SLA support? Lets talk.
Contact Us

AgentLayerWeb is licensed under the MIT License. Questions? Email hi@agentlayerweb.dev