algolia-data-handling

name	algolia-data-handling
description	Implement Algolia data handling: record transforms, PII filtering before indexing, data retention, GDPR/CCPA compliance with Algolia's deleteByQuery and Insights deletion. Trigger: "algolia data", "algolia PII", "algolia GDPR", "algolia data retention", "algolia privacy", "algolia CCPA", "algolia data sync".
allowed-tools	Read, Write, Edit
version	1.7.0
license	MIT
author	Jeremy Longshore <[e-mail]>
tags	["saas","search","algolia"]
compatibility	Designed for Claude Code

Algolia Data Handling

Overview

Algolia stores your records in their cloud. You control what data goes in (via saveObjects), what comes back (via attributesToRetrieve), and what users can search (via searchableAttributes). For privacy compliance, you must filter PII before indexing and implement deletion workflows.

Prerequisites

• A documented data classification for the source records and approval to index only the searchable fields.
• An Algolia Admin key held by the backend indexing job, never the browser.
• A deletion and retention owner for data-subject requests before production records are copied to Algolia.

Data Flow: Source → Algolia → User

Source Database        Transform           Algolia Index           Search Response
┌──────────┐     ┌──────────────┐     ┌──────────────────┐     ┌──────────────┐
│ Full user │     │ Strip PII    │     │ Searchable       │     │ Retrieved    │
│ record    │ ──▶ │ Truncate     │ ──▶ │ fields only      │ ──▶ │ fields only  │
│ (all cols)│     │ Normalize    │     │ + ranking data   │     │ (UI needs)   │
└──────────┘     └──────────────┘     └──────────────────┘     └──────────────┘

Instructions

Examples

The transformations below demonstrate an allowlist-first record shape, explicit PII removal, deletion handling, and bounded retention. Apply the same policy to every source connector rather than relying on dashboard settings alone.

Step 1: Transform Records Before Indexing

import { algoliasearch } from 'algoliasearch';

const client = algoliasearch(process.env.ALGOLIA_APP_ID!, process.env.ALGOLIA_ADMIN_KEY!);

// Define what goes into Algolia — NOT everything from your DB
interface AlgoliaProduct {
  objectID: string;
  name: string;
  description: string;     // Truncated, plain text
  category: ;
  : ;
  : ;
  : ;
  : ;
  : ;
  : [];
}

 ():  {
   {
    : dbRecord.,
    : dbRecord.,
    : (dbRecord.).(, ),
    : dbRecord.?. || ,
    : dbRecord.?. || ,
    : dbRecord. / ,
    : dbRecord. > ,
    : dbRecord.?.[]?. || ,
    : dbRecord. || ,
    : (dbRecord),
  };
}


 ():  {
   html?.(, ).(, ).() || ;
}

 (): [] {
   : [] = [];
   (record.) tags.();
   (record.) tags.();
   (record. > ) tags.();
   tags;
}