Getting Started

Overview

Hushh Agentic Developer APIs let you build consent-first personalization with A2A (Agent-to-Agent) JSON-RPC 2.0 workflows. Use the testable /api/a2a/{agent} endpoints in the docs to prototype, then move server-to-server for production workloads.

The story: from first signal to activated profile

Imagine a luxury retail app onboarding Ava. She shares her email, selects a few style preferences, and explicitly consents to profile enrichment. Your app then:

1. Creates a clean profile record in Supabase.
2. Enriches it with public and Gemini agents for deeper context.
3. Normalizes the profile into CRM-friendly likes, dislikes, and affinities.
4. Activates that profile in your CDP or personalization engine.

The result is a profile that is consented, current, and actionable.

What you will build

1. Create or query a profile in Supabase.
2. Enrich the profile with public and Gemini agents.
3. Normalize likes, dislikes, and brand affinities.
4. Activate the merged profile in your CRM or CDP.

What success looks like

• A single merged JSON profile that is ready for segmentation.
• A clear consent trail for every enrichment step.
• Faster personalization without manual data wrangling.

Prerequisites

• A Hushh developer account.
• An API key and session token.
• A minimal identity payload (email or phone).

Base URL

https://hushh.ai

Your first workflow (story-driven)

Use Ava as a reference profile. Start by enriching public data, then run Gemini for gap filling, and finally normalize the profile for brand activation.

Make your first request

curl -X POST "https://hushh.ai/api/a2a/public" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d "{\"text\":\"Provide a detailed JSON profile for Sundar Pichai, email sundar.pichai@example.com, phone +1 6505559001.\",\"sessionId\":\"session-123\",\"id\":\"task-abc123\"}"

Example identity payload

{
  "text": "Create a profile for Ava Torres with email ava.torres@example.com, phone +1 4155550199, city San Francisco, and interests luxury retail, travel, wellness.",
  "sessionId": "session-456",
  "id": "task-789"
}

Common pitfalls to avoid

• Using live endpoints from the browser. Use the doc endpoints for testing, then move to server-to-server.
• Skipping consent metadata before enrichment.
• Storing raw agent outputs without normalization.

Next steps

• Developer console setup
• Agentic endpoints
• Profile data and enrichment
• Use cases

Ready to explore? Start with the onboarding flow, then open the Agentic Endpoints page to test your first live request.