Typesense vs Algolia Search Cost & Crossover ROI

Copy// Drop-in Replacement for Algolia InstantSearch in React / Next.js
import instantsearch from 'instantsearch.js';
import { searchBox, hits } from 'instantsearch.js/es/widgets';
import TypesenseInstantSearchAdapter from 'typesense-instantsearch-adapter';

const typesenseInstantsearchAdapter = new TypesenseInstantSearchAdapter({
  server: {
    apiKey: 'your_search_only_api_key', // Search-only scoped key
    nodes: [
      {
        host: 'search.yourdomain.com', // Proxied via Cloudflare or direct
        port: 443,
        protocol: 'https',
      },
    ],
    cacheSearchResultsForSeconds: 120, // Client-side cache to save bandwidth
  },
  additionalSearchParameters: {
    query_by: 'name,description,category,brand',
    num_typos: 2,
  },
});

const searchClient = typesenseInstantsearchAdapter.searchClient;

// Exactly identical to Algolia instantsearch initialization!
const search = instantsearch({
  indexName: 'products',
  searchClient,
});

search.addWidgets([
  searchBox({ container: '#searchbox' }),
  hits({
    container: '#hits',
    templates: {
      item: (hit) => `<div><strong>${hit.name}</strong> - $${hit.price}</div>`,
    },
  }),
]);

search.start();

Copy# sync_pg_to_typesense.py - Bulk index products from PostgreSQL into Typesense
import psycopg2
import typesense

client = typesense.Client({
    'nodes': [{'host': 'localhost', 'port': '8108', 'protocol': 'http'}],
    'api_key': 'your_admin_api_key',
    'connection_timeout_seconds': 5
})

# Create Schema
schema = {
    'name': 'products',
    'fields': [
        {'name': 'id', 'type': 'string'},
        {'name': 'name', 'type': 'string'},
        {'name': 'description', 'type': 'string', 'optional': True},
        {'name': 'category', 'type': 'string', 'facet': True},
        {'name': 'price', 'type': 'float', 'facet': True}
    ]
}

try:
    client.collections['products'].delete()
except Exception:
    pass
client.collections.create(schema)

# Stream rows from PostgreSQL
conn = psycopg2.connect("dbname=ecommerce user=postgres host=localhost password=secret")
cur = conn.cursor(name='product_cursor')
cur.execute("SELECT id::text, name, description, category, price FROM products WHERE is_active = true")

batch = []
for row in cur:
    batch.append({
        'id': row[0],
        'name': row[1],
        'description': row[2] or '',
        'category': row[3] or 'General',
        'price': float(row[4])
    })
    if len(batch) >= 1000:
        client.collections['products'].documents.import_(batch, {'action': 'create'})
        batch = []

if batch:
    client.collections['products'].documents.import_(batch, {'action': 'create'})

print("[Success] All PostgreSQL products indexed into Typesense.")

Copy#!/bin/bash
# backup_typesense_to_r2.sh - Nightly automated snapshot to Cloudflare R2
DATE=$(date +%Y-%m-%d_%H%M%S)
BACKUP_DIR="/opt/typesense-backups/$DATE"
mkdir -p "$BACKUP_DIR"

# 1. Trigger Typesense Snapshot API
curl -X POST "http://localhost:8108/operations/snapshot?snapshot_path=/data/snapshot-$DATE"   -H "X-TYPESENSE-API-KEY: $TYPESENSE_API_KEY"

# 2. Archive and stream to Cloudflare R2 via rclone (Zero Egress!)
tar -czf "/tmp/typesense-backup-$DATE.tar.gz" -C /opt/typesense-data "snapshot-$DATE"
rclone copy "/tmp/typesense-backup-$DATE.tar.gz" "r2:search-backups/typesense/"

# 3. Clean local temporary snapshot
rm -rf "/opt/typesense-data/snapshot-$DATE" "/tmp/typesense-backup-$DATE.tar.gz"
echo "[Backup Complete] Streamed to Cloudflare R2: typesense-backup-$DATE.tar.gz"