World Agent Web

An ecosystem of protocols enabling AI agents to discover and interact with web services through standardized manifests

Think of it as: HTTP is to the World Wide Web what WAW is to the Agent Web

Three Core Protocols:
├─ HADP (HTTP Agent Discovery Protocol) — How agents find services
├─ ACDL (Agent Capability Description Language) — How services describe themselves
└─ AWCP (Agent-Web Communication Protocol) — How agents invoke services

The Problem Today

Currently, AI agents access web services through web search and HTML scraping — an approach designed for humans, not machines.

❌ Current Approach

  • Generic web search engines
  • Unstructured HTML scraping
  • Fragile (breaks when sites change)
  • No type safety or validation
  • Agent must guess capabilities
  • Hardcoded per-service integration

✅ WAW Solution

  • Standardized service discovery
  • JSON Schema validated data
  • Stable contracts (versioned)
  • Type-safe invocation
  • Self-describing capabilities
  • Protocol-based (no custom code)

Vision: The Agent Web

HTTP made the web accessible to humans. WAW makes the web accessible to AI agents.

1990s
Web 1.0

Static pages, read-only

2000s
Web 2.0

Interactive, user-generated content

2010s
Web 3.0

Semantic web, decentralized

2020s+
Agent Web

AI-native, standardized discovery

Today

Agents scrape HTML
meant for humans

Tomorrow

Services expose
agent-ready manifests

The Agent Protocols Ecosystem

WAW works alongside other protocols to enable complete agent infrastructure

┌─────────────────────────────────────────┐
│     Agent Orchestration Layer (A2A)     │
│   Agents collaborate on complex tasks   │
└─────────────────────────────────────────┘
↓             ↓              ↓
┌────────────┐ ┌────────────┐ ┌────────────┐
│    MCP     │ │    WAW     │ │   Custom   │
│  Desktop   │ │    Web     │ │    APIs    │
│   Tools    │ │  Services  │ │            │
└────────────┘ └────────────┘ └────────────┘
            

MCP

Scope: Desktop/Local

Transport: stdio, SSE

Use: IDE, terminal, local files

Best For: Desktop integration

WAW

Scope: Web/Cloud

Transport: HTTP/HTTPS

Use: APIs, SaaS, web services

Best For: Service invocation

A2A

Scope: Agent-to-Agent

Transport: Network

Use: Multi-agent collaboration

Best For: Orchestration

Detailed Comparison

How WAW relates to other approaches for agent-service integration

Feature Web Search MCP WAW A2A
Target Human browsing Desktop tools Web services Agent collaboration
Transport HTML stdio, SSE HTTP/HTTPS Network
Discovery Search engine Config file Manifest Registry
Data Format Unstructured JSON JSON Schema Protocol-specific
Best For General info IDE/Desktop apps APIs/SaaS Multi-agent tasks

Key Benefits

🔍
Discovery
Agents find services dynamically
Type Safety
JSON Schema validation built-in
🌐
Web Native
Standard HTTP/HTTPS transport
Orchestration
Chain multiple services seamlessly

Live Activity

Real-time usage statistics from production services

0
Total Requests
0
Bookings Created
0
Weather Queries

Features

Automatic Discovery

Services expose capabilities via standard manifest files. Agents discover and understand them without prior configuration.

Type Safety

JSON Schema validation ensures inputs and outputs conform to specifications. Catch errors before execution.

Web Native

Built on standard HTTP/HTTPS. Works with existing infrastructure. No special protocols or transports required.

Complementary Design

Works alongside MCP (desktop tools) and A2A (agent collaboration). Extends rather than replaces current solutions.

Simple Adoption

Add a manifest to your existing API. Deploy to any HTTP server. Services become instantly discoverable by agents.

Open Source

MIT licensed. Community-driven specification. Built for the open web, by the open web.

Future: The Agent Economy

As the Agent Web grows, new infrastructure and business models will emerge

🏛️

Service Registries

Like search engines for the web, registries will index and rank WAW services by quality, speed, and cost.

Inevitable
💳

x402 Micropayments

Agents pay for API calls automatically. No subscriptions, no invoicing—instant micropayments enable true pay-per-use.

Enabled
🏪

Service Marketplace

Businesses publish WAW-compliant services. Agents discover and use them. Revenue flows automatically through protocols.

Vision

Example: Agent Books Restaurant Autonomously

1. Discovery Agent queries registry: "booking services near Milan"
2. Selection Compares 10 services by rating, price (€0.01), latency
3. Payment x402 micropayment (€0.01) authorized instantly
4. Booking Service invoked, reservation confirmed

Quick Start

Agent Implementation
from src.client.client import WAWClient

# Initialize client
client = WAWClient()

# 1. Discover services
manifest = client.discover_manifest("https://web-production-73c2.up.railway.app")

# 2. Find weather service
weather = client.find_agent_by_id(manifest, "weather-service")

# 3. Invoke capability
response = client.invoke_capability(
    agent=weather,
    capability_id="current-weather",
    params={"city": "Milan", "units": "metric"}
)

print(response['result'])
# Output: {'city': 'Milan', 'temperature': 18.5, ...}
Service Implementation
# 1. Create manifest describing your service
{
  "version": "1.0",
  "agents": [{
    "id": "my-service",
    "capabilities": [{
      "id": "process",
      "input_schema": { "type": "object", ... },
      "output_schema": { "type": "object", ... }
    }]
  }]
}

# 2. Serve at /agent-manifest.json
# 3. Agents can now discover and use your service

Try the Full Implementation

Explore complete examples, autonomous agent demos, and production-ready code on GitHub.

Autonomous agent demo with Claude
Complete client library with validation
Working examples (Weather, Booking, Calculator)
Testing scripts and API reference
View on GitHub →

Available Services

Live demo services accessible through the WAW protocols

🌤️
Weather Service

Current conditions and forecasts for 15+ cities worldwide

🍽️
Booking Service

Search, check availability, and book at 25+ restaurants

🔢
Calculator Service

Mathematical expressions and statistical analysis

🔊
Echo Service

Test service for protocols validation