Docs

Operating notes for real production usage.

CleanCSV AI is designed for quick cleanup before imports, reporting, and analytics handoffs. This page documents what is automated, what still requires review, and what to monitor before going live.

Storage mode
Persistent

Redis-backed deployments keep report sessions across restarts. Local or unconfigured environments keep sessions in memory for about one hour.

Billing path
Creem

Checkout, restoration, and customer portal management all flow throughCreem. Webhook delivery is required for durable subscription state.

Ops endpoint
/api/health

Returns deployment readiness for uploads, billing, webhooks, Redis, and optional AI suggestions. Use it in uptime monitoring.

Supported cleanup scope

Uploads accept CSV, TSV, XLSX, and XLS files.

Rule-based cleanup covers duplicates, blanks, mixed date formats, encoding artifacts, numeric formatting, whitespace, and safe split-column helpers.

Non-deterministic cases such as invalid emails, type mismatches, and outliers stay review-only unless a future rule can prove a safe change.

Recommended production checklist

1. Set `NEXT_PUBLIC_URL`, the active billing provider keys, the matching webhook secret, and Upstash Redis.

2. Verify `/api/health` returns `productionReady: true` after deploy.

3. Run a live Creem test payment and confirm portal access from `/pricing`.

4. Upload a real customer sample and verify preview, fixes, and export format.