Neon

Gotchas

• 100 CU-hours/month per project

  Each free project gets 100 compute unit-hours per calendar month; compute auto-scales to zero when idle so bursts are the primary consumer.

• 0.5 GB storage per project

  Each free-tier project is capped at 0.5 GB of database storage, which is very tight for anything beyond toy datasets.

• Scale-to-zero after 5 min idle

  Free-tier computes automatically suspend after 5 minutes of inactivity and incur a cold-start latency (~350 ms) on the next connection.

• 6-hour time travel / restore window

  Point-in-time restore is limited to the last 6 hours (or 1 GB of WAL changes); paid plans extend this to 7–30 days.

• Max 2 CU compute size (8 GB RAM)

  Free computes can scale up to only 2 CU (≈1 vCPU / 8 GB RAM); paid plans reach 16–56 CU.

• 10 branches per project

  Each free project is limited to 10 concurrent branches; extra branches require a paid plan.

• 1-day monitoring / log retention

  Metrics and logs in the Neon console are retained for only 1 day on the free tier vs. 3–14 days on paid plans.

• 1 snapshot only

  Free tier includes just 1 manual snapshot; Launch and Scale plans include 100 snapshots with scheduled backup options.

• 60K MAU cap on Neon Auth

  The built-in authentication feature (Neon Auth) is capped at 60,000 monthly active users on the free plan.

• 5 GB/month egress included

  Public network transfer (egress) is capped at 5 GB per month on the free tier; overages require upgrading.

• No IP allow rules or private networking

  IP allowlisting and private network (PrivateLink) access are Scale-plan-only features, unavailable on free.

• No SLA, SOC 2, or HIPAA compliance

  Free tier has no uptime SLA and no access to SOC 2 or HIPAA compliance reports; these are Scale-plan features.

• Community support only (Discord)

  Free accounts receive community Discord support only; billing and dedicated support require a paid plan.