example4.ai — Real code examples for AI agents

example4.ai

Real, buildable code examples for AI agents.
Stop burning tokens on documentation.

Join the waitlist See how it works ↓

Launching at AI Week Italia · May 2026

The problem

Documentation is a token tax.

🔥

Tokens burned to generate it

AI spends thousands of tokens synthesizing documentation that goes stale in weeks.

🔥

Tokens burned to read it

Your AI agent parses pages of prose just to figure out how to call a method.

❌

Never verifiable

No CI can tell you if documentation is still correct. Code can.

What if your AI could learn from real code instead?

The solution

Code that documents itself.

example4.ai is an MCP server that gives your AI agent real, buildable, testable code examples for any library. No prose. No markdown. Just code that works.

If it builds, it's true. If tests pass, it's correct.

{
 "library": "Newtonsoft.Json",
 "version": "13.0.3",
 "examples": [
   {
     "name": "basic-serialization",
     "complexity": "beginner",
     "build": "dotnet build",
     "test": "dotnet test"
   }
 ]
}

How it works

Three steps. Zero docs.

Plug in.

Connect example4.ai as an MCP server to your AI agent.

{
  "mcpServers": {
    "example4": {
      "url": "https://example4.ai/mcp"
    }
  }
}

Ask.

Your AI asks how to use a library.

// Your prompt:

"How do I use custom
 JsonConverters in
 Newtonsoft.Json?"

Get real code.

Receives working code from a project that builds and passes tests.

public class DateConverter
  : JsonConverter<DateTime>
{
  public override DateTime
    ReadJson(JsonReader r, ...)
  {
    var s = (string)r.Value;
    return DateTime.Parse(s);
  }
}

Comparison

example4.ai vs the old way.

	Traditional Docs	DeepWiki	example4.ai
Format	Prose / Markdown	Generated wiki	Buildable code
Verifiable	❌	❌	✅ CI/CD
Freshness	❌ Stale	❌ Snapshot	✅ Push-triggered
Token cost	🔥🔥🔥	🔥🔥	🟢 Minimal
AI-native	❌	❌	✅ MCP + LSAI
Executable	❌	❌	✅ dotnet test

Ecosystem

Multi-language from day one.

C# TypeScript Python Java Go Rust

And growing with the community.

Community

Publish your examples.
Grow the ecosystem.

example4.ai is like npm for code examples. Publish example repos for your library, automatically verified by CI, and make them available to every AI agent on the planet.

Library maintainers

Increase adoption of your library.

Developers

Contribute examples, build reputation.

AI Agents

Access verified code via MCP.

Semantic code navigation that understands types, relationships, and call graphs. Not just grep.

Learn more about LSAI →

Be the first to kill your docs.

Join the waitlist for early access. Launching May 2026.

You're in.

We'll notify you when example4.ai goes live.