AMP
GitHub
Getting Started

Quick Start

Integrate AMP into your AI agent in 5 minutes.

1. Get Your API Key

Sign up and get your API key from the AMP dashboard.

💡 Tip: Keep your API key secure. Never commit it to version control.

2. Install the SDK

Choose your language:

TypeScript / JavaScript

npm install @amp-protocol/client

Python

pip install amp-protocol

3. Initialise the Client

agent.ts
import { AMP } from '@amp-protocol/client';

const amp = new AMP({
  apiKey: process.env.AMP_API_KEY,
});

4. Get Context Before Responding

Before your agent responds to a user request, call getContext():

agent.ts
// User asks: "Help me build a login page"

const context = await amp.getContext({
  userId: user.id,
  task: "build a login page",
  metadata: {
    timestamp: Date.now(),
    source: "chat",
  },
});

// AMP returns:
// {
//   requestId: "req_abc123",
//   suggestedFraming: "micro_task",
//   communicationStyle: "brief_directive",
//   complexity: "break_into_steps",
//   effectiveness: 0.87
// }

5. Adapt Your Response

Use the context to personalise your agent's behaviour:

agent.ts
// Option 1: Adapt your system prompt
const systemPrompt = `You are a coding assistant.
Use ${context.communicationStyle} communication.
Frame this task as a ${context.suggestedFraming}.`;

const response = await openai.chat.completions.create({
  model: "gpt-4",
  messages: [
    { role: "system", content: systemPrompt },
    { role: "user", content: userQuery },
  ],
});

// Option 2: Adapt programmatically
if (context.complexity === "break_into_steps") {
  response = generateStepByStepGuide(userQuery);
} else {
  response = generateDirectSolution(userQuery);
}

6. Report the Outcome

After the user interacts with your response, report what happened:

agent.ts
await amp.reportOutcome({
  requestId: context.requestId,
  started: true,          // Did the user start the task?
  completed: true,        // Did they complete it?
  timeToStart: 45,        // Seconds until they started
  flowState: true,        // Did they seem engaged?
  satisfaction: 0.9,      // Optional: user feedback score
});

Complete Example

Here's a full integration example:

agent.ts
import { AMP } from '@amp-protocol/client';
import OpenAI from 'openai';

const amp = new AMP({ apiKey: process.env.AMP_API_KEY });
const openai = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });

export async function handleUserQuery(
  userId: string,
  query: string
): Promise<string> {
  // 1. Get motivation context
  const context = await amp.getContext({
    userId,
    task: query,
    metadata: { timestamp: Date.now() },
  });

  // 2. Adapt your system prompt
  const systemPrompt = `You are a helpful coding assistant.
Communication style: ${context.communicationStyle}
Task framing: ${context.suggestedFraming}
Complexity approach: ${context.complexity}`;

  // 3. Generate response
  const response = await openai.chat.completions.create({
    model: "gpt-4",
    messages: [
      { role: "system", content: systemPrompt },
      { role: "user", content: query },
    ],
  });

  const answer = response.choices[0].message.content;

  // 4. Report outcome (in background or after user interaction)
  setTimeout(async () => {
    await amp.reportOutcome({
      requestId: context.requestId,
      started: true,
      completed: true,
    });
  }, 30000);

  return answer;
}

Next Steps